id sid tid token lemma pos bracis-33583 1 1 growing grow VERB bracis-33583 1 2 self self NOUN bracis-33583 1 3 - - PUNCT bracis-33583 1 4 organizing organize VERB bracis-33583 1 5 maps map NOUN bracis-33583 1 6 for for ADP bracis-33583 1 7     SPACE bracis-33583 1 8 multi multi ADJ bracis-33583 1 9 - - ADJ bracis-33583 1 10 label label ADJ bracis-33583 1 11 classification classification NOUN bracis-33583 1 12 | | NOUN bracis-33583 1 13 springer springer NOUN bracis-33583 1 14 nature nature PROPN bracis-33583 1 15 link link PROPN bracis-33583 1 16 ( ( PUNCT bracis-33583 1 17 formerly formerly ADV bracis-33583 1 18 springerlink springerlink NOUN bracis-33583 1 19 ) ) PUNCT bracis-33583 1 20 skip skip VERB bracis-33583 1 21 to to ADP bracis-33583 1 22 main main ADJ bracis-33583 1 23 content content NOUN bracis-33583 1 24 advertisement advertisement NOUN bracis-33583 1 25 log log NOUN bracis-33583 1 26 in in ADP bracis-33583 1 27 menu menu NOUN bracis-33583 1 28 find find VERB bracis-33583 1 29 a a DET bracis-33583 1 30 journal journal NOUN bracis-33583 1 31 publish publish VERB bracis-33583 1 32 with with ADP bracis-33583 1 33 us we PRON bracis-33583 1 34 track track VERB bracis-33583 1 35 your your PRON bracis-33583 1 36 research research NOUN bracis-33583 1 37 search search NOUN bracis-33583 1 38 cart cart NOUN bracis-33583 1 39 home home NOUN bracis-33583 1 40 intelligent intelligent ADJ bracis-33583 1 41 systems system NOUN bracis-33583 1 42 conference conference NOUN bracis-33583 1 43 paper paper NOUN bracis-33583 1 44 growing grow VERB bracis-33583 1 45 self self NOUN bracis-33583 1 46 - - PUNCT bracis-33583 1 47 organizing organize VERB bracis-33583 1 48 maps map NOUN bracis-33583 1 49 for for ADP bracis-33583 1 50     SPACE bracis-33583 1 51 multi multi ADJ bracis-33583 1 52 - - ADJ bracis-33583 1 53 label label ADJ bracis-33583 1 54 classification classification NOUN bracis-33583 1 55 conference conference NOUN bracis-33583 1 56 paper paper NOUN bracis-33583 1 57 first first ADV bracis-33583 1 58 online online ADV bracis-33583 1 59 : : PUNCT bracis-33583 1 60 30 30 NUM bracis-33583 1 61 january january NOUN bracis-33583 1 62 2025 2025 NUM bracis-33583 1 63 pp pp ADP bracis-33583 1 64 33–48 33–48 NUM bracis-33583 1 65 cite cite VERB bracis-33583 1 66 this this DET bracis-33583 1 67 conference conference NOUN bracis-33583 1 68 paper paper NOUN bracis-33583 1 69 access access NOUN bracis-33583 1 70 provided provide VERB bracis-33583 1 71 by by ADP bracis-33583 1 72 university university PROPN bracis-33583 1 73 of of ADP bracis-33583 1 74 notre notre PROPN bracis-33583 1 75 dame dame PROPN bracis-33583 1 76 hesburgh hesburgh PROPN bracis-33583 1 77 library library PROPN bracis-33583 1 78 download download PROPN bracis-33583 1 79 book book NOUN bracis-33583 1 80 pdf pdf PROPN bracis-33583 1 81 download download NOUN bracis-33583 1 82 book book NOUN bracis-33583 1 83 epub epub PROPN bracis-33583 1 84 intelligent intelligent ADJ bracis-33583 1 85 systems system NOUN bracis-33583 1 86 ( ( PUNCT bracis-33583 1 87 bracis bracis NOUN bracis-33583 1 88 2024 2024 NUM bracis-33583 1 89 ) ) PUNCT bracis-33583 1 90 growing grow VERB bracis-33583 1 91 self self NOUN bracis-33583 1 92 - - PUNCT bracis-33583 1 93 organizing organize VERB bracis-33583 1 94 maps map NOUN bracis-33583 1 95 for for ADP bracis-33583 1 96     SPACE bracis-33583 1 97 multi multi ADJ bracis-33583 1 98 - - ADJ bracis-33583 1 99 label label ADJ bracis-33583 1 100 classification classification NOUN bracis-33583 1 101 download download NOUN bracis-33583 1 102 book book NOUN bracis-33583 1 103 pdf pdf PROPN bracis-33583 1 104 download download NOUN bracis-33583 1 105 book book NOUN bracis-33583 1 106 epub epub PROPN bracis-33583 1 107 pedro pedro PROPN bracis-33583 1 108 henrique henrique PROPN bracis-33583 1 109 casarotto9 casarotto9 PROPN bracis-33583 1 110 & & CCONJ bracis-33583 1 111 ricardo ricardo PROPN bracis-33583 1 112 cerri cerri PROPN bracis-33583 1 113     SPACE bracis-33583 1 114 orcid orcid NOUN bracis-33583 1 115 : : PUNCT bracis-33583 1 116 orcid.org/0000-0002-2582-169510 orcid.org/0000-0002-2582-169510 VERB bracis-33583 1 117     SPACE bracis-33583 1 118 part part NOUN bracis-33583 1 119 of of ADP bracis-33583 1 120 the the DET bracis-33583 1 121 book book NOUN bracis-33583 1 122 series series NOUN bracis-33583 1 123 : : PUNCT bracis-33583 1 124 lecture lecture NOUN bracis-33583 1 125 notes note NOUN bracis-33583 1 126 in in ADP bracis-33583 1 127 computer computer NOUN bracis-33583 1 128 science science NOUN bracis-33583 1 129 ( ( PUNCT bracis-33583 1 130 ( ( PUNCT bracis-33583 1 131 lnai lnai ADJ bracis-33583 1 132 , , PUNCT bracis-33583 1 133 volume volume NOUN bracis-33583 1 134 15413 15413 NUM bracis-33583 1 135 ) ) PUNCT bracis-33583 1 136 ) ) PUNCT bracis-33583 1 137 included include VERB bracis-33583 1 138 in in ADP bracis-33583 1 139 the the DET bracis-33583 1 140 following follow VERB bracis-33583 1 141 conference conference NOUN bracis-33583 1 142 series series NOUN bracis-33583 1 143 : : PUNCT bracis-33583 1 144 brazilian brazilian ADJ bracis-33583 1 145 conference conference NOUN bracis-33583 1 146 on on ADP bracis-33583 1 147 intelligent intelligent ADJ bracis-33583 1 148 systems system NOUN bracis-33583 1 149 399 399 NUM bracis-33583 1 150 accesses access NOUN bracis-33583 1 151 1 1 NUM bracis-33583 1 152 citation citation NOUN bracis-33583 1 153 abstract abstract NOUN bracis-33583 1 154 in in ADP bracis-33583 1 155 machine machine NOUN bracis-33583 1 156 learning learning NOUN bracis-33583 1 157 , , PUNCT bracis-33583 1 158 multi multi ADJ bracis-33583 1 159 - - ADJ bracis-33583 1 160 label label ADJ bracis-33583 1 161 classification classification NOUN bracis-33583 1 162 is be AUX bracis-33583 1 163 the the DET bracis-33583 1 164 problem problem NOUN bracis-33583 1 165 of of ADP bracis-33583 1 166 simultaneously simultaneously ADV bracis-33583 1 167 classifying classify VERB bracis-33583 1 168 an an DET bracis-33583 1 169 instance instance NOUN bracis-33583 1 170 into into ADP bracis-33583 1 171 two two NUM bracis-33583 1 172 or or CCONJ bracis-33583 1 173 more more ADJ bracis-33583 1 174 labels label NOUN bracis-33583 1 175 . . PUNCT bracis-33583 2 1 it it PRON bracis-33583 2 2 is be AUX bracis-33583 2 3 a a DET bracis-33583 2 4 challenging challenging ADJ bracis-33583 2 5 problem problem NOUN bracis-33583 2 6 since since SCONJ bracis-33583 2 7 each each DET bracis-33583 2 8 label label NOUN bracis-33583 2 9 has have VERB bracis-33583 2 10 its its PRON bracis-33583 2 11 specialty specialty NOUN bracis-33583 2 12 , , PUNCT bracis-33583 2 13 and and CCONJ bracis-33583 2 14 correlations correlation NOUN bracis-33583 2 15 between between ADP bracis-33583 2 16 them they PRON bracis-33583 2 17 must must AUX bracis-33583 2 18 be be AUX bracis-33583 2 19 considered consider VERB bracis-33583 2 20 . . PUNCT bracis-33583 3 1 a a DET bracis-33583 3 2 self self NOUN bracis-33583 3 3 - - PUNCT bracis-33583 3 4 organizing organize VERB bracis-33583 3 5 map map NOUN bracis-33583 3 6 ( ( PUNCT bracis-33583 3 7 som som NOUN bracis-33583 3 8 ) ) PUNCT bracis-33583 3 9 is be AUX bracis-33583 3 10 a a DET bracis-33583 3 11 neural neural ADJ bracis-33583 3 12 network network NOUN bracis-33583 3 13 where where SCONJ bracis-33583 3 14 neurons neuron NOUN bracis-33583 3 15 organized organize VERB bracis-33583 3 16 in in ADP bracis-33583 3 17 a a DET bracis-33583 3 18 grid grid NOUN bracis-33583 3 19 are be AUX bracis-33583 3 20 tuned tune VERB bracis-33583 3 21 to to PART bracis-33583 3 22 represent represent VERB bracis-33583 3 23 the the DET bracis-33583 3 24 input input NOUN bracis-33583 3 25 instances instance NOUN bracis-33583 3 26 in in ADP bracis-33583 3 27 self self NOUN bracis-33583 3 28 - - PUNCT bracis-33583 3 29 organization organization NOUN bracis-33583 3 30 . . PUNCT bracis-33583 4 1 after after ADP bracis-33583 4 2 tuning tune VERB bracis-33583 4 3 , , PUNCT bracis-33583 4 4 similar similar ADJ bracis-33583 4 5 instances instance NOUN bracis-33583 4 6 in in ADP bracis-33583 4 7 the the DET bracis-33583 4 8 input input NOUN bracis-33583 4 9 space space NOUN bracis-33583 4 10 are be AUX bracis-33583 4 11 mapped map VERB bracis-33583 4 12 to to ADP bracis-33583 4 13 closer close ADJ bracis-33583 4 14 neurons neuron NOUN bracis-33583 4 15 in in ADP bracis-33583 4 16 the the DET bracis-33583 4 17 grid grid NOUN bracis-33583 4 18 . . PUNCT bracis-33583 5 1 soms som NOUN bracis-33583 5 2 have have AUX bracis-33583 5 3 already already ADV bracis-33583 5 4 been be AUX bracis-33583 5 5 used use VERB bracis-33583 5 6 for for ADP bracis-33583 5 7 multi multi ADJ bracis-33583 5 8 - - ADJ bracis-33583 5 9 label label ADJ bracis-33583 5 10 problems problem NOUN bracis-33583 5 11 , , PUNCT bracis-33583 5 12 obtaining obtain VERB bracis-33583 5 13 competitive competitive ADJ bracis-33583 5 14 results result NOUN bracis-33583 5 15 with with ADP bracis-33583 5 16 other other ADJ bracis-33583 5 17 methods method NOUN bracis-33583 5 18 . . PUNCT bracis-33583 6 1 however however ADV bracis-33583 6 2 , , PUNCT bracis-33583 6 3 the the DET bracis-33583 6 4 static static ADJ bracis-33583 6 5 nature nature NOUN bracis-33583 6 6 of of ADP bracis-33583 6 7 their their PRON bracis-33583 6 8 grid grid NOUN bracis-33583 6 9 of of ADP bracis-33583 6 10 neurons neuron NOUN bracis-33583 6 11 is be AUX bracis-33583 6 12 a a DET bracis-33583 6 13 disadvantage disadvantage NOUN bracis-33583 6 14 since since SCONJ bracis-33583 6 15 it it PRON bracis-33583 6 16 is be AUX bracis-33583 6 17 difficult difficult ADJ bracis-33583 6 18 to to PART bracis-33583 6 19 define define VERB bracis-33583 6 20 the the DET bracis-33583 6 21 optimized optimize VERB bracis-33583 6 22 grid grid NOUN bracis-33583 6 23 size size NOUN bracis-33583 6 24 for for ADP bracis-33583 6 25 each each DET bracis-33583 6 26 problem problem NOUN bracis-33583 6 27 . . PUNCT bracis-33583 7 1 the the DET bracis-33583 7 2 growing grow VERB bracis-33583 7 3 self self NOUN bracis-33583 7 4 - - PUNCT bracis-33583 7 5 organizing organize VERB bracis-33583 7 6 maps map NOUN bracis-33583 7 7 ( ( PUNCT bracis-33583 7 8 gsom gsom PROPN bracis-33583 7 9 ) ) PUNCT bracis-33583 7 10 extends extend VERB bracis-33583 7 11 the the DET bracis-33583 7 12 soms som NOUN bracis-33583 7 13 , , PUNCT bracis-33583 7 14 allowing allow VERB bracis-33583 7 15 the the DET bracis-33583 7 16 network network NOUN bracis-33583 7 17 to to PART bracis-33583 7 18 grow grow VERB bracis-33583 7 19 during during ADP bracis-33583 7 20 execution execution NOUN bracis-33583 7 21 based base VERB bracis-33583 7 22 on on ADP bracis-33583 7 23 the the DET bracis-33583 7 24 data data NOUN bracis-33583 7 25 characteristics characteristic NOUN bracis-33583 7 26 . . PUNCT bracis-33583 8 1 this this DET bracis-33583 8 2 paper paper NOUN bracis-33583 8 3 proposes propose VERB bracis-33583 8 4 a a DET bracis-33583 8 5 gsom gsom NOUN bracis-33583 8 6 to to PART bracis-33583 8 7 predict predict VERB bracis-33583 8 8 multi multi ADJ bracis-33583 8 9 - - ADJ bracis-33583 8 10 label label ADJ bracis-33583 8 11 data datum NOUN bracis-33583 8 12 . . PUNCT bracis-33583 9 1 the the DET bracis-33583 9 2 experiments experiment NOUN bracis-33583 9 3 showed show VERB bracis-33583 9 4 that that PRON bracis-33583 9 5 gsom gsom PROPN bracis-33583 9 6 obtained obtain VERB bracis-33583 9 7 better well ADV bracis-33583 9 8 or or CCONJ bracis-33583 9 9 more more ADV bracis-33583 9 10 competitive competitive ADJ bracis-33583 9 11 results result NOUN bracis-33583 9 12 in in ADP bracis-33583 9 13 most most ADJ bracis-33583 9 14 of of ADP bracis-33583 9 15 the the DET bracis-33583 9 16 datasets dataset NOUN bracis-33583 9 17 investigated investigate VERB bracis-33583 9 18 compared compare VERB bracis-33583 9 19 to to ADP bracis-33583 9 20 som som VERB bracis-33583 9 21 and and CCONJ bracis-33583 9 22 had have VERB bracis-33583 9 23 a a DET bracis-33583 9 24 competitive competitive ADJ bracis-33583 9 25 performance performance NOUN bracis-33583 9 26 compared compare VERB bracis-33583 9 27 to to ADP bracis-33583 9 28 other other ADJ bracis-33583 9 29 methods method NOUN bracis-33583 9 30 . . PUNCT bracis-33583 10 1 access access NOUN bracis-33583 10 2 provided provide VERB bracis-33583 10 3 by by ADP bracis-33583 10 4 university university PROPN bracis-33583 10 5 of of ADP bracis-33583 10 6 notre notre PROPN bracis-33583 10 7 dame dame PROPN bracis-33583 10 8 hesburgh hesburgh PROPN bracis-33583 10 9 library library PROPN bracis-33583 10 10 . . PUNCT bracis-33583 11 1 download download PROPN bracis-33583 11 2 conference conference NOUN bracis-33583 11 3 paper paper NOUN bracis-33583 11 4 pdf pdf NOUN bracis-33583 11 5 similar similar ADJ bracis-33583 11 6 content content NOUN bracis-33583 11 7 being be AUX bracis-33583 11 8 viewed view VERB bracis-33583 11 9 by by ADP bracis-33583 11 10 others other NOUN bracis-33583 11 11 incorporation incorporation NOUN bracis-33583 11 12 of of ADP bracis-33583 11 13 neighborhood neighborhood NOUN bracis-33583 11 14 concept concept NOUN bracis-33583 11 15     SPACE bracis-33583 11 16 in in ADP bracis-33583 11 17 enhancing enhance VERB bracis-33583 11 18 som som PROPN bracis-33583 11 19 based base VERB bracis-33583 11 20 multi multi ADJ bracis-33583 11 21 - - ADJ bracis-33583 11 22 label label ADJ bracis-33583 11 23 classification classification NOUN bracis-33583 11 24 chapter chapter NOUN bracis-33583 11 25 © © PROPN bracis-33583 11 26 2019 2019 NUM bracis-33583 11 27 improving improve VERB bracis-33583 11 28 self self NOUN bracis-33583 11 29 - - PUNCT bracis-33583 11 30 organizing organize VERB bracis-33583 11 31 maps map NOUN bracis-33583 11 32 with with ADP bracis-33583 11 33 unsupervised unsupervised ADJ bracis-33583 11 34 feature feature NOUN bracis-33583 11 35 extraction extraction NOUN bracis-33583 11 36 chapter chapter NOUN bracis-33583 11 37 © © NOUN bracis-33583 11 38 2020 2020 NUM bracis-33583 11 39 enhancing enhance VERB bracis-33583 11 40 visual visual ADJ bracis-33583 11 41 clustering clustering NOUN bracis-33583 11 42 using use VERB bracis-33583 11 43 adaptive adaptive ADJ bracis-33583 11 44 moving move VERB bracis-33583 11 45 self self NOUN bracis-33583 11 46 - - PUNCT bracis-33583 11 47 organizing organize VERB bracis-33583 11 48 maps map NOUN bracis-33583 11 49 ( ( PUNCT bracis-33583 11 50 amsom amsom NOUN bracis-33583 11 51 ) ) PUNCT bracis-33583 11 52 chapter chapter NOUN bracis-33583 11 53 © © NOUN bracis-33583 11 54 2017 2017 NUM bracis-33583 11 55 1 1 NUM bracis-33583 11 56 introduction introduction NOUN bracis-33583 11 57 in in ADP bracis-33583 11 58 the the DET bracis-33583 11 59 machine machine NOUN bracis-33583 11 60 learning learn VERB bracis-33583 11 61 context context NOUN bracis-33583 11 62 , , PUNCT bracis-33583 11 63 the the DET bracis-33583 11 64 most most ADV bracis-33583 11 65 common common ADJ bracis-33583 11 66 classification classification NOUN bracis-33583 11 67 problem problem NOUN bracis-33583 11 68 is be AUX bracis-33583 11 69 called call VERB bracis-33583 11 70 single single ADJ bracis-33583 11 71 - - PUNCT bracis-33583 11 72 label label NOUN bracis-33583 11 73 . . PUNCT bracis-33583 12 1 this this DET bracis-33583 12 2 problem problem NOUN bracis-33583 12 3 involves involve VERB bracis-33583 12 4 classifying classify VERB bracis-33583 12 5 a a DET bracis-33583 12 6 data data NOUN bracis-33583 12 7 input input NOUN bracis-33583 12 8 in in ADP bracis-33583 12 9 a a DET bracis-33583 12 10 single single ADJ bracis-33583 12 11 class class NOUN bracis-33583 12 12 l l NOUN bracis-33583 12 13 from from ADP bracis-33583 12 14 a a DET bracis-33583 12 15 set set NOUN bracis-33583 12 16 of of ADP bracis-33583 12 17 classes class NOUN bracis-33583 12 18 l l NOUN bracis-33583 12 19 , , PUNCT bracis-33583 12 20 where where SCONJ bracis-33583 12 21 \(|l| \(|l| X bracis-33583 12 22 > > X bracis-33583 12 23 1\ 1\ NUM bracis-33583 12 24 ) ) PUNCT bracis-33583 12 25 . . PUNCT bracis-33583 13 1 if if SCONJ bracis-33583 13 2 the the DET bracis-33583 13 3 data data NOUN bracis-33583 13 4 input input NOUN bracis-33583 13 5 is be AUX bracis-33583 13 6 expected expect VERB bracis-33583 13 7 to to PART bracis-33583 13 8 be be AUX bracis-33583 13 9 classified classify VERB bracis-33583 13 10 over over ADP bracis-33583 13 11 two two NUM bracis-33583 13 12 or or CCONJ bracis-33583 13 13 more more ADJ bracis-33583 13 14 classes class NOUN bracis-33583 13 15 simultaneously simultaneously ADV bracis-33583 13 16 , , PUNCT bracis-33583 13 17 then then ADV bracis-33583 13 18 the the DET bracis-33583 13 19 problem problem NOUN bracis-33583 13 20 is be AUX bracis-33583 13 21 considered consider VERB bracis-33583 13 22 a a DET bracis-33583 13 23 multi multi ADJ bracis-33583 13 24 - - ADJ bracis-33583 13 25 label label ADJ bracis-33583 13 26 classification classification NOUN bracis-33583 13 27 problem problem NOUN bracis-33583 13 28 . . PUNCT bracis-33583 14 1 in in ADP bracis-33583 14 2 the the DET bracis-33583 14 3 literature literature NOUN bracis-33583 14 4 , , PUNCT bracis-33583 14 5 this this DET bracis-33583 14 6 problem problem NOUN bracis-33583 14 7 is be AUX bracis-33583 14 8 presented present VERB bracis-33583 14 9 over over ADP bracis-33583 14 10 diverse diverse ADJ bracis-33583 14 11 areas area NOUN bracis-33583 14 12 , , PUNCT bracis-33583 14 13 such such ADJ bracis-33583 14 14 image image NOUN bracis-33583 14 15 classification classification NOUN bracis-33583 14 16 and and CCONJ bracis-33583 14 17 recognition recognition NOUN bracis-33583 14 18     SPACE bracis-33583 15 1 [ [ X bracis-33583 15 2 8 8 NUM bracis-33583 15 3 , , PUNCT bracis-33583 15 4 22 22 NUM bracis-33583 15 5 ] ] PUNCT bracis-33583 15 6 , , PUNCT bracis-33583 15 7 text text NOUN bracis-33583 15 8 and and CCONJ bracis-33583 15 9 document document NOUN bracis-33583 15 10 classification classification NOUN bracis-33583 15 11     SPACE bracis-33583 16 1 [ [ X bracis-33583 16 2 9 9 NUM bracis-33583 16 3 , , PUNCT bracis-33583 16 4 12 12 NUM bracis-33583 16 5 ] ] PUNCT bracis-33583 16 6 , , PUNCT bracis-33583 16 7 music music NOUN bracis-33583 16 8 classification classification NOUN bracis-33583 16 9     SPACE bracis-33583 17 1 [ [ X bracis-33583 17 2 33 33 NUM bracis-33583 17 3 , , PUNCT bracis-33583 17 4 40 40 NUM bracis-33583 17 5 ] ] PUNCT bracis-33583 17 6 , , PUNCT bracis-33583 17 7 in in ADP bracis-33583 17 8 biomedicine biomedicine NOUN bracis-33583 17 9 , , PUNCT bracis-33583 17 10 such such ADJ bracis-33583 17 11 in in ADP bracis-33583 17 12 genome genome NOUN bracis-33583 17 13 classification classification NOUN bracis-33583 17 14     SPACE bracis-33583 18 1 [ [ X bracis-33583 18 2 39 39 NUM bracis-33583 18 3 , , PUNCT bracis-33583 18 4 42 42 NUM bracis-33583 18 5 ] ] PUNCT bracis-33583 18 6 and and CCONJ bracis-33583 18 7 protein protein NOUN bracis-33583 18 8 function function NOUN bracis-33583 18 9 prediction prediction NOUN bracis-33583 18 10     SPACE bracis-33583 19 1 [ [ X bracis-33583 19 2 29 29 NUM bracis-33583 19 3 , , PUNCT bracis-33583 19 4 30 30 NUM bracis-33583 19 5 ] ] PUNCT bracis-33583 19 6 . . PUNCT bracis-33583 20 1 there there PRON bracis-33583 20 2 are be VERB bracis-33583 20 3 many many ADJ bracis-33583 20 4 strategies strategy NOUN bracis-33583 20 5 to to PART bracis-33583 20 6 deal deal VERB bracis-33583 20 7 with with ADP bracis-33583 20 8 multi multi ADJ bracis-33583 20 9 - - ADJ bracis-33583 20 10 label label ADJ bracis-33583 20 11 problems problem NOUN bracis-33583 20 12 , , PUNCT bracis-33583 20 13 divided divide VERB bracis-33583 20 14 into into ADP bracis-33583 20 15 two two NUM bracis-33583 20 16 main main ADJ bracis-33583 20 17 approaches approach NOUN bracis-33583 20 18 . . PUNCT bracis-33583 21 1 the the DET bracis-33583 21 2 first first ADJ bracis-33583 21 3 is be AUX bracis-33583 21 4 called call VERB bracis-33583 21 5 the the DET bracis-33583 21 6 local local ADJ bracis-33583 21 7 approach approach NOUN bracis-33583 21 8 , , PUNCT bracis-33583 21 9 in in ADP bracis-33583 21 10 which which PRON bracis-33583 21 11 traditional traditional ADJ bracis-33583 21 12 classification classification NOUN bracis-33583 21 13 algorithms algorithm NOUN bracis-33583 21 14 independently independently ADV bracis-33583 21 15 predict predict VERB bracis-33583 21 16 each each DET bracis-33583 21 17 label label NOUN bracis-33583 21 18 . . PUNCT bracis-33583 22 1 the the DET bracis-33583 22 2 other other ADJ bracis-33583 22 3 strategy strategy NOUN bracis-33583 22 4 is be AUX bracis-33583 22 5 global global ADJ bracis-33583 22 6 , , PUNCT bracis-33583 22 7 which which PRON bracis-33583 22 8 considers consider VERB bracis-33583 22 9 label label NOUN bracis-33583 22 10 correlations correlation NOUN bracis-33583 22 11 by by ADP bracis-33583 22 12 creating create VERB bracis-33583 22 13 specific specific ADJ bracis-33583 22 14 algorithms algorithm NOUN bracis-33583 22 15 to to PART bracis-33583 22 16 deal deal VERB bracis-33583 22 17 with with ADP bracis-33583 22 18 the the DET bracis-33583 22 19 problem problem NOUN bracis-33583 22 20 and and CCONJ bracis-33583 22 21 all all DET bracis-33583 22 22 labels label NOUN bracis-33583 22 23 simultaneously simultaneously ADV bracis-33583 22 24 . . PUNCT bracis-33583 23 1 a a DET bracis-33583 23 2 hybrid hybrid NOUN bracis-33583 23 3 of of ADP bracis-33583 23 4 these these DET bracis-33583 23 5 two two NUM bracis-33583 23 6 approaches approach NOUN bracis-33583 23 7 is be AUX bracis-33583 23 8 also also ADV bracis-33583 23 9 possible possible ADJ bracis-33583 23 10     SPACE bracis-33583 23 11 [ [ X bracis-33583 23 12 19 19 NUM bracis-33583 23 13 ] ] PUNCT bracis-33583 23 14 . . PUNCT bracis-33583 24 1 colombini colombini PROPN bracis-33583 24 2 et et PROPN bracis-33583 24 3 al al PROPN bracis-33583 24 4 . . PUNCT bracis-33583 24 5     SPACE bracis-33583 25 1 [ [ X bracis-33583 25 2 17 17 NUM bracis-33583 25 3 ] ] PUNCT bracis-33583 25 4 proposed propose VERB bracis-33583 25 5 using use VERB bracis-33583 25 6 self self NOUN bracis-33583 25 7 - - PUNCT bracis-33583 25 8 organizing organize VERB bracis-33583 25 9 maps map NOUN bracis-33583 25 10     SPACE bracis-33583 26 1 [ [ X bracis-33583 26 2 21 21 NUM bracis-33583 26 3 ] ] PUNCT bracis-33583 26 4 to to PART bracis-33583 26 5 solve solve VERB bracis-33583 26 6 the the DET bracis-33583 26 7 multi multi ADJ bracis-33583 26 8 - - ADJ bracis-33583 26 9 label label ADJ bracis-33583 26 10 problems problem NOUN bracis-33583 26 11 . . PUNCT bracis-33583 27 1 self self NOUN bracis-33583 27 2 - - PUNCT bracis-33583 27 3 organizing organize VERB bracis-33583 27 4 maps map NOUN bracis-33583 27 5 ( ( PUNCT bracis-33583 27 6 som som NOUN bracis-33583 27 7 ) ) PUNCT bracis-33583 27 8 , , PUNCT bracis-33583 27 9 or or CCONJ bracis-33583 27 10 kohonen kohonen PROPN bracis-33583 27 11 maps map NOUN bracis-33583 27 12 , , PUNCT bracis-33583 27 13 is be AUX bracis-33583 27 14 an an DET bracis-33583 27 15 unsupervised unsupervised ADJ bracis-33583 27 16 neural neural ADJ bracis-33583 27 17 network network NOUN bracis-33583 27 18 model model NOUN bracis-33583 27 19 containing contain VERB bracis-33583 27 20 a a DET bracis-33583 27 21 one one NUM bracis-33583 27 22 - - PUNCT bracis-33583 27 23 layer layer NOUN bracis-33583 27 24 grid grid NOUN bracis-33583 27 25 of of ADP bracis-33583 27 26 neurons neuron NOUN bracis-33583 27 27 . . PUNCT bracis-33583 28 1 each each DET bracis-33583 28 2 neuron neuron NOUN bracis-33583 28 3 has have VERB bracis-33583 28 4 a a DET bracis-33583 28 5 discriminant discriminant NOUN bracis-33583 28 6 function function NOUN bracis-33583 28 7 measuring measure VERB bracis-33583 28 8 the the DET bracis-33583 28 9 similarity similarity NOUN bracis-33583 28 10 of of ADP bracis-33583 28 11 the the DET bracis-33583 28 12 neuron neuron NOUN bracis-33583 28 13 weights weight VERB bracis-33583 28 14 to to ADP bracis-33583 28 15 an an DET bracis-33583 28 16 input input NOUN bracis-33583 28 17 instance instance NOUN bracis-33583 28 18 . . PUNCT bracis-33583 29 1 in in ADP bracis-33583 29 2 addition addition NOUN bracis-33583 29 3 , , PUNCT bracis-33583 29 4 colombini colombini PROPN bracis-33583 29 5 et et PROPN bracis-33583 29 6 al al PROPN bracis-33583 29 7 . . PROPN bracis-33583 29 8 associated associate VERB bracis-33583 29 9 a a DET bracis-33583 29 10 prototype prototype NOUN bracis-33583 29 11 vector vector NOUN bracis-33583 29 12 to to ADP bracis-33583 29 13 each each DET bracis-33583 29 14 neuron neuron NOUN bracis-33583 29 15 , , PUNCT bracis-33583 29 16 which which PRON bracis-33583 29 17 is be AUX bracis-33583 29 18 the the DET bracis-33583 29 19 mean mean NOUN bracis-33583 29 20 of of ADP bracis-33583 29 21 the the DET bracis-33583 29 22 input input NOUN bracis-33583 29 23 label label NOUN bracis-33583 29 24 vectors vector NOUN bracis-33583 29 25 of of ADP bracis-33583 29 26 the the DET bracis-33583 29 27 instances instance NOUN bracis-33583 29 28 mapped map VERB bracis-33583 29 29 to to ADP bracis-33583 29 30 the the DET bracis-33583 29 31 neuron neuron NOUN bracis-33583 29 32 . . PUNCT bracis-33583 30 1 when when SCONJ bracis-33583 30 2 the the DET bracis-33583 30 3 model model NOUN bracis-33583 30 4 receives receive VERB bracis-33583 30 5 a a DET bracis-33583 30 6 new new ADJ bracis-33583 30 7 instance instance NOUN bracis-33583 30 8 , , PUNCT bracis-33583 30 9 it it PRON bracis-33583 30 10 is be AUX bracis-33583 30 11 compared compare VERB bracis-33583 30 12 with with ADP bracis-33583 30 13 all all DET bracis-33583 30 14 weights weight NOUN bracis-33583 30 15 of of ADP bracis-33583 30 16 the the DET bracis-33583 30 17 neurons neuron NOUN bracis-33583 30 18 of of ADP bracis-33583 30 19 the the DET bracis-33583 30 20 grid grid NOUN bracis-33583 30 21 . . PUNCT bracis-33583 31 1 the the DET bracis-33583 31 2 neuron neuron NOUN bracis-33583 31 3 that that PRON bracis-33583 31 4 shows show VERB bracis-33583 31 5 more more ADJ bracis-33583 31 6 similarity similarity NOUN bracis-33583 31 7 with with ADP bracis-33583 31 8 the the DET bracis-33583 31 9 instance instance NOUN bracis-33583 31 10 is be AUX bracis-33583 31 11 selected select VERB bracis-33583 31 12 as as ADP bracis-33583 31 13 the the DET bracis-33583 31 14 winner winner NOUN bracis-33583 31 15 neuron neuron NOUN bracis-33583 31 16 , , PUNCT bracis-33583 31 17 and and CCONJ bracis-33583 31 18 his his PRON bracis-33583 31 19 prototype prototype NOUN bracis-33583 31 20 vector vector NOUN bracis-33583 31 21 is be AUX bracis-33583 31 22 returned return VERB bracis-33583 31 23 as as ADP bracis-33583 31 24 the the DET bracis-33583 31 25 prediction prediction NOUN bracis-33583 31 26 for for ADP bracis-33583 31 27 that that DET bracis-33583 31 28 instance instance NOUN bracis-33583 31 29 . . PUNCT bracis-33583 32 1 the the DET bracis-33583 32 2 growing grow VERB bracis-33583 32 3 self self NOUN bracis-33583 32 4 - - PUNCT bracis-33583 32 5 organizing organize VERB bracis-33583 32 6 maps map NOUN bracis-33583 32 7 ( ( PUNCT bracis-33583 32 8 gsom gsom PROPN bracis-33583 32 9 ) ) PUNCT bracis-33583 32 10     SPACE bracis-33583 33 1 [ [ X bracis-33583 33 2 6 6 NUM bracis-33583 33 3 ] ] PUNCT bracis-33583 33 4 are be AUX bracis-33583 33 5 an an DET bracis-33583 33 6 extension extension NOUN bracis-33583 33 7 of of ADP bracis-33583 33 8 soms som NOUN bracis-33583 33 9 in in ADP bracis-33583 33 10 which which PRON bracis-33583 33 11 the the DET bracis-33583 33 12 number number NOUN bracis-33583 33 13 of of ADP bracis-33583 33 14 neurons neuron NOUN bracis-33583 33 15 in in ADP bracis-33583 33 16 the the DET bracis-33583 33 17 grid grid NOUN bracis-33583 33 18 increases increase NOUN bracis-33583 33 19 during during ADP bracis-33583 33 20 training training NOUN bracis-33583 33 21 . . PUNCT bracis-33583 34 1 the the DET bracis-33583 34 2 increase increase NOUN bracis-33583 34 3 rate rate NOUN bracis-33583 34 4 is be AUX bracis-33583 34 5 established establish VERB bracis-33583 34 6 by by ADP bracis-33583 34 7 the the DET bracis-33583 34 8 growth growth NOUN bracis-33583 34 9 threshold threshold NOUN bracis-33583 34 10 , , PUNCT bracis-33583 34 11 a a DET bracis-33583 34 12 parameter parameter NOUN bracis-33583 34 13 that that PRON bracis-33583 34 14 compares compare VERB bracis-33583 34 15 the the DET bracis-33583 34 16 total total ADJ bracis-33583 34 17 error error NOUN bracis-33583 34 18 of of ADP bracis-33583 34 19 the the DET bracis-33583 34 20 neuron neuron NOUN bracis-33583 34 21 with with ADP bracis-33583 34 22 the the DET bracis-33583 34 23 input input NOUN bracis-33583 34 24 data datum NOUN bracis-33583 34 25 . . PUNCT bracis-33583 35 1 a a DET bracis-33583 35 2 new new ADJ bracis-33583 35 3 neuron neuron NOUN bracis-33583 35 4 is be AUX bracis-33583 35 5 created create VERB bracis-33583 35 6 if if SCONJ bracis-33583 35 7 the the DET bracis-33583 35 8 error error NOUN bracis-33583 35 9 value value NOUN bracis-33583 35 10 exceeds exceed VERB bracis-33583 35 11 the the DET bracis-33583 35 12 growth growth NOUN bracis-33583 35 13 threshold threshold NOUN bracis-33583 35 14 . . PUNCT bracis-33583 36 1 the the DET bracis-33583 36 2 benefit benefit NOUN bracis-33583 36 3 of of ADP bracis-33583 36 4 this this DET bracis-33583 36 5 adaptation adaptation NOUN bracis-33583 36 6 is be AUX bracis-33583 36 7 that that SCONJ bracis-33583 36 8 the the DET bracis-33583 36 9 number number NOUN bracis-33583 36 10 of of ADP bracis-33583 36 11 neurons neuron NOUN bracis-33583 36 12 is be AUX bracis-33583 36 13 more more ADV bracis-33583 36 14 precise precise ADJ bracis-33583 36 15 , , PUNCT bracis-33583 36 16 and and CCONJ bracis-33583 36 17 thus thus ADV bracis-33583 36 18 , , PUNCT bracis-33583 36 19 the the DET bracis-33583 36 20 results result NOUN bracis-33583 36 21 are be AUX bracis-33583 36 22 better well ADJ bracis-33583 36 23 . . PUNCT bracis-33583 37 1 in in ADP bracis-33583 37 2 this this DET bracis-33583 37 3 paper paper NOUN bracis-33583 37 4 , , PUNCT bracis-33583 37 5 we we PRON bracis-33583 37 6 propose propose VERB bracis-33583 37 7 a a DET bracis-33583 37 8 method method NOUN bracis-33583 37 9 for for ADP bracis-33583 37 10 multi multi ADJ bracis-33583 37 11 - - ADJ bracis-33583 37 12 label label ADJ bracis-33583 37 13 classification classification NOUN bracis-33583 37 14 using use VERB bracis-33583 37 15 growing grow VERB bracis-33583 37 16 self self NOUN bracis-33583 37 17 - - PUNCT bracis-33583 37 18 organizing organize VERB bracis-33583 37 19 maps map NOUN bracis-33583 37 20 . . PUNCT bracis-33583 38 1 the the DET bracis-33583 38 2 method method NOUN bracis-33583 38 3 is be AUX bracis-33583 38 4 called call VERB bracis-33583 38 5 growing grow VERB bracis-33583 38 6 self self NOUN bracis-33583 38 7 - - PUNCT bracis-33583 38 8 organizing organize VERB bracis-33583 38 9 maps map NOUN bracis-33583 38 10 multi multi ADJ bracis-33583 38 11 - - ADJ bracis-33583 38 12 label label ADJ bracis-33583 38 13 learning learning NOUN bracis-33583 38 14 ( ( PUNCT bracis-33583 38 15 gsom gsom PROPN bracis-33583 38 16 - - PUNCT bracis-33583 38 17 mll mll PROPN bracis-33583 38 18 ) ) PUNCT bracis-33583 38 19 , , PUNCT bracis-33583 38 20 and and CCONJ bracis-33583 38 21 improves improve VERB bracis-33583 38 22 the the DET bracis-33583 38 23 som som PROPN bracis-33583 38 24 - - PUNCT bracis-33583 38 25 mll mll PROPN bracis-33583 38 26 proposed propose VERB bracis-33583 38 27 in in ADP bracis-33583 38 28 the the DET bracis-33583 38 29 work work NOUN bracis-33583 38 30 of of ADP bracis-33583 38 31 colombini colombini PROPN bracis-33583 38 32 et et PROPN bracis-33583 38 33 al al PROPN bracis-33583 38 34 . . PUNCT bracis-33583 38 35     SPACE bracis-33583 39 1 [ [ X bracis-33583 39 2 17 17 NUM bracis-33583 39 3 ] ] PUNCT bracis-33583 39 4 . . PUNCT bracis-33583 40 1 the the DET bracis-33583 40 2 remainder remainder NOUN bracis-33583 40 3 of of ADP bracis-33583 40 4 this this DET bracis-33583 40 5 paper paper NOUN bracis-33583 40 6 is be AUX bracis-33583 40 7 organized organize VERB bracis-33583 40 8 as as SCONJ bracis-33583 40 9 follows follow VERB bracis-33583 40 10 . . PUNCT bracis-33583 41 1 section section NOUN bracis-33583 41 2     SPACE bracis-33583 41 3 2 2 NUM bracis-33583 41 4 contains contain VERB bracis-33583 41 5 a a DET bracis-33583 41 6 literature literature NOUN bracis-33583 41 7 review review NOUN bracis-33583 41 8 of of ADP bracis-33583 41 9 the the DET bracis-33583 41 10 most most ADV bracis-33583 41 11 common common ADJ bracis-33583 41 12 algorithms algorithm NOUN bracis-33583 41 13 for for ADP bracis-33583 41 14 multi multi ADJ bracis-33583 41 15 - - ADJ bracis-33583 41 16 label label ADJ bracis-33583 41 17 problems problem NOUN bracis-33583 41 18 . . PUNCT bracis-33583 42 1 section section NOUN bracis-33583 42 2     SPACE bracis-33583 42 3 3 3 NUM bracis-33583 42 4 presents present VERB bracis-33583 42 5 an an DET bracis-33583 42 6 overview overview NOUN bracis-33583 42 7 of of ADP bracis-33583 42 8 the the DET bracis-33583 42 9 kohonen kohonen PROPN bracis-33583 42 10 maps maps PROPN bracis-33583 42 11 and and CCONJ bracis-33583 42 12 introduces introduce VERB bracis-33583 42 13 the the DET bracis-33583 42 14 som som PROPN bracis-33583 42 15 - - PUNCT bracis-33583 42 16 mll mll PROPN bracis-33583 42 17 proposed propose VERB bracis-33583 42 18 by by ADP bracis-33583 42 19 colombini colombini PROPN bracis-33583 42 20 et et PROPN bracis-33583 42 21 al al PROPN bracis-33583 42 22 . . PUNCT bracis-33583 42 23     SPACE bracis-33583 43 1 [ [ X bracis-33583 43 2 17 17 NUM bracis-33583 43 3 ] ] PUNCT bracis-33583 43 4 . . PUNCT bracis-33583 44 1 our our PRON bracis-33583 44 2 proposal proposal NOUN bracis-33583 44 3 , , PUNCT bracis-33583 44 4 gsom gsom PROPN bracis-33583 44 5 - - PUNCT bracis-33583 44 6 mll mll PROPN bracis-33583 44 7 , , PUNCT bracis-33583 44 8 is be AUX bracis-33583 44 9 presented present VERB bracis-33583 44 10 in in ADP bracis-33583 44 11 sect sect NOUN bracis-33583 44 12 . . PUNCT bracis-33583 44 13     SPACE bracis-33583 45 1 4 4 NUM bracis-33583 45 2 , , PUNCT bracis-33583 45 3 while while SCONJ bracis-33583 45 4 our our PRON bracis-33583 45 5 methodology methodology NOUN bracis-33583 45 6 is be AUX bracis-33583 45 7 presented present VERB bracis-33583 45 8 in in ADP bracis-33583 45 9 sect sect NOUN bracis-33583 45 10 . . PUNCT bracis-33583 45 11     SPACE bracis-33583 46 1 5 5 NUM bracis-33583 46 2 . . PUNCT bracis-33583 46 3 section section NOUN bracis-33583 46 4     SPACE bracis-33583 46 5 6 6 NUM bracis-33583 46 6 presents present NOUN bracis-33583 46 7 and and CCONJ bracis-33583 46 8 discusses discuss VERB bracis-33583 46 9 our our PRON bracis-33583 46 10 experiments experiment NOUN bracis-33583 46 11 and and CCONJ bracis-33583 46 12 results result NOUN bracis-33583 46 13 . . PUNCT bracis-33583 47 1 finally finally ADV bracis-33583 47 2 , , PUNCT bracis-33583 47 3 sect sect NOUN bracis-33583 47 4 . . PUNCT bracis-33583 47 5     SPACE bracis-33583 47 6 7 7 NUM bracis-33583 47 7 concludes conclude VERB bracis-33583 47 8 the the DET bracis-33583 47 9 paper paper NOUN bracis-33583 47 10 and and CCONJ bracis-33583 47 11 points point NOUN bracis-33583 47 12 to to ADP bracis-33583 47 13 future future ADJ bracis-33583 47 14 work work NOUN bracis-33583 47 15 . . PUNCT bracis-33583 48 1 2 2 NUM bracis-33583 48 2 multi multi ADJ bracis-33583 48 3 - - ADJ bracis-33583 48 4 label label ADJ bracis-33583 48 5 classification classification NOUN bracis-33583 48 6 this this DET bracis-33583 48 7 section section NOUN bracis-33583 48 8 presents present VERB bracis-33583 48 9 the the DET bracis-33583 48 10 two two NUM bracis-33583 48 11 main main ADJ bracis-33583 48 12 approaches approach NOUN bracis-33583 48 13 to to ADP bracis-33583 48 14 classifying classify VERB bracis-33583 48 15 multi multi ADJ bracis-33583 48 16 - - ADJ bracis-33583 48 17 label label ADJ bracis-33583 48 18 data datum NOUN bracis-33583 48 19 : : PUNCT bracis-33583 48 20 local local ADJ bracis-33583 48 21 ( ( PUNCT bracis-33583 48 22 problem problem NOUN bracis-33583 48 23 transformation transformation NOUN bracis-33583 48 24 ) ) PUNCT bracis-33583 48 25 and and CCONJ bracis-33583 48 26 global global ADJ bracis-33583 48 27 ( ( PUNCT bracis-33583 48 28 algorithm algorithm NOUN bracis-33583 48 29 adaptation adaptation NOUN bracis-33583 48 30 ) ) PUNCT bracis-33583 48 31 . . PUNCT bracis-33583 49 1 2.1 2.1 NUM bracis-33583 49 2 local local ADJ bracis-33583 49 3 approach approach NOUN bracis-33583 49 4 the the DET bracis-33583 49 5 local local ADJ bracis-33583 49 6 approach approach NOUN bracis-33583 49 7 reduces reduce VERB bracis-33583 49 8 the the DET bracis-33583 49 9 original original ADJ bracis-33583 49 10 problem problem NOUN bracis-33583 49 11 into into ADP bracis-33583 49 12 many many ADJ bracis-33583 49 13 different different ADJ bracis-33583 49 14 single single ADJ bracis-33583 49 15 - - PUNCT bracis-33583 49 16 label label NOUN bracis-33583 49 17 problems problem NOUN bracis-33583 49 18 . . PUNCT bracis-33583 50 1 each each DET bracis-33583 50 2 problem problem NOUN bracis-33583 50 3 is be AUX bracis-33583 50 4 individually individually ADV bracis-33583 50 5 solved solve VERB bracis-33583 50 6 , , PUNCT bracis-33583 50 7 and and CCONJ bracis-33583 50 8 the the DET bracis-33583 50 9 corresponding corresponding ADJ bracis-33583 50 10 predictions prediction NOUN bracis-33583 50 11 are be AUX bracis-33583 50 12 combined combine VERB bracis-33583 50 13 to to PART bracis-33583 50 14 form form VERB bracis-33583 50 15 the the DET bracis-33583 50 16 final final ADJ bracis-33583 50 17 multi multi ADJ bracis-33583 50 18 - - ADJ bracis-33583 50 19 label label ADJ bracis-33583 50 20 classification classification NOUN bracis-33583 50 21 . . PUNCT bracis-33583 51 1 the the DET bracis-33583 51 2 simplest simple ADJ bracis-33583 51 3 local local ADJ bracis-33583 51 4 - - PUNCT bracis-33583 51 5 based base VERB bracis-33583 51 6 strategy strategy NOUN bracis-33583 51 7 is be AUX bracis-33583 51 8 called call VERB bracis-33583 51 9 binary binary ADJ bracis-33583 51 10 - - PUNCT bracis-33583 51 11 relevance relevance NOUN bracis-33583 51 12 ( ( PUNCT bracis-33583 51 13 br br NOUN bracis-33583 51 14 ) ) PUNCT bracis-33583 51 15     SPACE bracis-33583 52 1 [ [ X bracis-33583 52 2 35 35 NUM bracis-33583 52 3 ] ] PUNCT bracis-33583 52 4 in in ADP bracis-33583 52 5 which which PRON bracis-33583 52 6 l l NOUN bracis-33583 52 7 binary binary ADJ bracis-33583 52 8 classifiers classifier NOUN bracis-33583 52 9 are be AUX bracis-33583 52 10 trained train VERB bracis-33583 52 11 , , PUNCT bracis-33583 52 12 each each PRON bracis-33583 52 13 associated associate VERB bracis-33583 52 14 with with ADP bracis-33583 52 15 one one NUM bracis-33583 52 16 of of ADP bracis-33583 52 17 the the DET bracis-33583 52 18 l l NOUN bracis-33583 52 19 labels label NOUN bracis-33583 52 20 . . PUNCT bracis-33583 53 1 one one NUM bracis-33583 53 2 advantage advantage NOUN bracis-33583 53 3 of of ADP bracis-33583 53 4 this this DET bracis-33583 53 5 strategy strategy NOUN bracis-33583 53 6 is be AUX bracis-33583 53 7 that that SCONJ bracis-33583 53 8 the the DET bracis-33583 53 9 classifiers classifier NOUN bracis-33583 53 10 can can AUX bracis-33583 53 11 be be AUX bracis-33583 53 12 run run VERB bracis-33583 53 13 independently independently ADV bracis-33583 53 14 from from ADP bracis-33583 53 15 each each DET bracis-33583 53 16 other other ADJ bracis-33583 53 17 ( ( PUNCT bracis-33583 53 18 in in ADP bracis-33583 53 19 parallel parallel NOUN bracis-33583 53 20 , , PUNCT bracis-33583 53 21 for for ADP bracis-33583 53 22 example example NOUN bracis-33583 53 23 ) ) PUNCT bracis-33583 53 24 , , PUNCT bracis-33583 53 25 and and CCONJ bracis-33583 53 26 the the DET bracis-33583 53 27 peculiarities peculiarity NOUN bracis-33583 53 28 of of ADP bracis-33583 53 29 each each DET bracis-33583 53 30 label label NOUN bracis-33583 53 31 independently independently ADV bracis-33583 53 32 influence influence VERB bracis-33583 53 33 its its PRON bracis-33583 53 34 corresponding corresponding ADJ bracis-33583 53 35 classifier classifier NOUN bracis-33583 53 36 . . PUNCT bracis-33583 54 1 another another DET bracis-33583 54 2 advantage advantage NOUN bracis-33583 54 3 is be AUX bracis-33583 54 4 that that SCONJ bracis-33583 54 5 traditional traditional ADJ bracis-33583 54 6 algorithms algorithm NOUN bracis-33583 54 7 can can AUX bracis-33583 54 8 be be AUX bracis-33583 54 9 used use VERB bracis-33583 54 10 . . PUNCT bracis-33583 55 1 however however ADV bracis-33583 55 2 , , PUNCT bracis-33583 55 3 the the DET bracis-33583 55 4 disadvantage disadvantage NOUN bracis-33583 55 5 is be AUX bracis-33583 55 6 that that SCONJ bracis-33583 55 7 this this DET bracis-33583 55 8 strategy strategy NOUN bracis-33583 55 9 ignores ignore VERB bracis-33583 55 10 the the DET bracis-33583 55 11 correlations correlation NOUN bracis-33583 55 12 between between ADP bracis-33583 55 13 labels label NOUN bracis-33583 55 14 , , PUNCT bracis-33583 55 15 a a DET bracis-33583 55 16 fundamental fundamental ADJ bracis-33583 55 17 characteristic characteristic NOUN bracis-33583 55 18 of of ADP bracis-33583 55 19 multi multi ADJ bracis-33583 55 20 - - ADJ bracis-33583 55 21 label label ADJ bracis-33583 55 22 problems problem NOUN bracis-33583 55 23 . . PUNCT bracis-33583 56 1 many many ADJ bracis-33583 56 2 proposed propose VERB bracis-33583 56 3 improvements improvement NOUN bracis-33583 56 4 in in ADP bracis-33583 56 5 the the DET bracis-33583 56 6 br br PROPN bracis-33583 56 7 algorithm algorithm NOUN bracis-33583 56 8 have have AUX bracis-33583 56 9 been be AUX bracis-33583 56 10 explored explore VERB bracis-33583 56 11 over over ADP bracis-33583 56 12 the the DET bracis-33583 56 13 last last ADJ bracis-33583 56 14 few few ADJ bracis-33583 56 15 years year NOUN bracis-33583 56 16 . . PUNCT bracis-33583 57 1 classifier classifier NOUN bracis-33583 57 2 chains chain NOUN bracis-33583 57 3 ( ( PUNCT bracis-33583 57 4 cc cc NOUN bracis-33583 57 5 ) ) PUNCT bracis-33583 57 6     SPACE bracis-33583 58 1 [ [ X bracis-33583 58 2 27 27 NUM bracis-33583 58 3 ] ] PUNCT bracis-33583 58 4 modifies modify VERB bracis-33583 58 5 br br PROPN bracis-33583 58 6 , , PUNCT bracis-33583 58 7 adding add VERB bracis-33583 58 8 a a DET bracis-33583 58 9 dependency dependency NOUN bracis-33583 58 10 mechanism mechanism NOUN bracis-33583 58 11 in in ADP bracis-33583 58 12 the the DET bracis-33583 58 13 labels label NOUN bracis-33583 58 14 . . PUNCT bracis-33583 59 1 this this PRON bracis-33583 59 2 is be AUX bracis-33583 59 3 done do VERB bracis-33583 59 4 by by ADP bracis-33583 59 5 a a DET bracis-33583 59 6 chain chain NOUN bracis-33583 59 7 that that PRON bracis-33583 59 8 randomly randomly ADV bracis-33583 59 9 connects connect VERB bracis-33583 59 10 the the DET bracis-33583 59 11 binary binary ADJ bracis-33583 59 12 classifiers classifier NOUN bracis-33583 59 13 , , PUNCT bracis-33583 59 14 considering consider VERB bracis-33583 59 15 the the DET bracis-33583 59 16 previous previous ADJ bracis-33583 59 17 labels label NOUN bracis-33583 59 18 / / SYM bracis-33583 59 19 predictions prediction NOUN bracis-33583 59 20 as as ADP bracis-33583 59 21 input input NOUN bracis-33583 59 22 for for ADP bracis-33583 59 23 training train VERB bracis-33583 59 24 a a DET bracis-33583 59 25 classifier classifier NOUN bracis-33583 59 26 in in ADP bracis-33583 59 27 the the DET bracis-33583 59 28 chain chain NOUN bracis-33583 59 29 . . PUNCT bracis-33583 60 1 it it PRON bracis-33583 60 2 has have VERB bracis-33583 60 3 the the DET bracis-33583 60 4 advantage advantage NOUN bracis-33583 60 5 of of ADP bracis-33583 60 6 maintaining maintain VERB bracis-33583 60 7 br br PROPN bracis-33583 60 8 ’s ’s PART bracis-33583 60 9 simplicity simplicity NOUN bracis-33583 60 10 while while SCONJ bracis-33583 60 11 learning learn VERB bracis-33583 60 12 the the DET bracis-33583 60 13 labels label NOUN bracis-33583 60 14 ’ ' PUNCT bracis-33583 60 15 relationships relationship NOUN bracis-33583 60 16 . . PUNCT bracis-33583 61 1 a a DET bracis-33583 61 2 similar similar ADJ bracis-33583 61 3 strategy strategy NOUN bracis-33583 61 4 is be AUX bracis-33583 61 5 meta meta X bracis-33583 61 6 staking staking NOUN bracis-33583 61 7 ( ( PUNCT bracis-33583 61 8 ms ms NOUN bracis-33583 61 9 ) ) PUNCT bracis-33583 61 10     SPACE bracis-33583 62 1 [ [ X bracis-33583 62 2 20 20 NUM bracis-33583 62 3 ] ] PUNCT bracis-33583 62 4 , , PUNCT bracis-33583 62 5 which which PRON bracis-33583 62 6 indirectly indirectly ADV bracis-33583 62 7 considers consider VERB bracis-33583 62 8 these these DET bracis-33583 62 9 relations relation NOUN bracis-33583 62 10 by by ADP bracis-33583 62 11 adding add VERB bracis-33583 62 12 the the DET bracis-33583 62 13 number number NOUN bracis-33583 62 14 of of ADP bracis-33583 62 15 examples example NOUN bracis-33583 62 16 based base VERB bracis-33583 62 17 on on ADP bracis-33583 62 18 the the DET bracis-33583 62 19 previous previous ADJ bracis-33583 62 20 classification classification NOUN bracis-33583 62 21 . . PUNCT bracis-33583 63 1 in in ADP bracis-33583 63 2 cherman cherman PROPN bracis-33583 63 3 et et PROPN bracis-33583 63 4 al al PROPN bracis-33583 63 5 . . PUNCT bracis-33583 63 6     SPACE bracis-33583 63 7 [ [ X bracis-33583 63 8 15 15 NUM bracis-33583 63 9 ] ] PUNCT bracis-33583 63 10 , , PUNCT bracis-33583 63 11 br br PROPN bracis-33583 63 12 was be AUX bracis-33583 63 13 explored explore VERB bracis-33583 63 14 using use VERB bracis-33583 63 15 decision decision NOUN bracis-33583 63 16 trees tree NOUN bracis-33583 63 17 and and CCONJ bracis-33583 63 18 naïve naïve ADJ bracis-33583 63 19 bayes bayes NOUN bracis-33583 63 20 to to PART bracis-33583 63 21 consider consider VERB bracis-33583 63 22 label label NOUN bracis-33583 63 23 relations relation NOUN bracis-33583 63 24 . . PUNCT bracis-33583 64 1 dembczyński dembczyński PROPN bracis-33583 64 2 et et PROPN bracis-33583 64 3 al al PROPN bracis-33583 64 4 . . PUNCT bracis-33583 64 5     SPACE bracis-33583 65 1 [ [ X bracis-33583 65 2 14 14 NUM bracis-33583 65 3 ] ] PUNCT bracis-33583 65 4 is be AUX bracis-33583 65 5 another another DET bracis-33583 65 6 approach approach NOUN bracis-33583 65 7 that that PRON bracis-33583 65 8 also also ADV bracis-33583 65 9 uses use VERB bracis-33583 65 10 the the DET bracis-33583 65 11 naïve naïve ADJ bracis-33583 65 12 bayes bayes NOUN bracis-33583 65 13 algorithm algorithm NOUN bracis-33583 65 14 to to PART bracis-33583 65 15 optimize optimize VERB bracis-33583 65 16 predictions prediction NOUN bracis-33583 65 17 . . PUNCT bracis-33583 66 1 the the DET bracis-33583 66 2 brknn brknn PROPN bracis-33583 66 3     SPACE bracis-33583 67 1 [ [ X bracis-33583 67 2 32 32 NUM bracis-33583 67 3 ] ] PUNCT bracis-33583 67 4 is be AUX bracis-33583 67 5 another another DET bracis-33583 67 6 approach approach NOUN bracis-33583 67 7 that that PRON bracis-33583 67 8 uses use VERB bracis-33583 67 9 the the DET bracis-33583 67 10 knn knn PROPN bracis-33583 67 11 algorithm algorithm NOUN bracis-33583 67 12 conceptually conceptually ADV bracis-33583 67 13 equivalent equivalent ADJ bracis-33583 67 14 to to ADP bracis-33583 67 15 br br PROPN bracis-33583 67 16 , , PUNCT bracis-33583 67 17 with with ADP bracis-33583 67 18 the the DET bracis-33583 67 19 advantage advantage NOUN bracis-33583 67 20 of of ADP bracis-33583 67 21 being be AUX bracis-33583 67 22 faster fast ADJ bracis-33583 67 23 . . PUNCT bracis-33583 68 1 another another DET bracis-33583 68 2 common common ADJ bracis-33583 68 3 local local ADJ bracis-33583 68 4 - - PUNCT bracis-33583 68 5 based base VERB bracis-33583 68 6 strategy strategy NOUN bracis-33583 68 7 is be AUX bracis-33583 68 8 called call VERB bracis-33583 68 9 label label NOUN bracis-33583 68 10 - - PUNCT bracis-33583 68 11 powerset powerset NOUN bracis-33583 68 12 ( ( PUNCT bracis-33583 68 13 lp lp NOUN bracis-33583 68 14 ) ) PUNCT bracis-33583 68 15     SPACE bracis-33583 69 1 [ [ X bracis-33583 69 2 11 11 NUM bracis-33583 69 3 ] ] PUNCT bracis-33583 69 4 in in ADP bracis-33583 69 5 which which PRON bracis-33583 69 6 all all DET bracis-33583 69 7 classes class NOUN bracis-33583 69 8 assigned assign VERB bracis-33583 69 9 to to ADP bracis-33583 69 10 each each DET bracis-33583 69 11 instance instance NOUN bracis-33583 69 12 are be AUX bracis-33583 69 13 turned turn VERB bracis-33583 69 14 into into ADP bracis-33583 69 15 a a DET bracis-33583 69 16 new new ADJ bracis-33583 69 17 and and CCONJ bracis-33583 69 18 unique unique ADJ bracis-33583 69 19 class class NOUN bracis-33583 69 20 . . PUNCT bracis-33583 70 1 this this PRON bracis-33583 70 2 is be AUX bracis-33583 70 3 a a DET bracis-33583 70 4 form form NOUN bracis-33583 70 5 to to PART bracis-33583 70 6 consider consider VERB bracis-33583 70 7 the the DET bracis-33583 70 8 correlations correlation NOUN bracis-33583 70 9 between between ADP bracis-33583 70 10 classes class NOUN bracis-33583 70 11 , , PUNCT bracis-33583 70 12 something something PRON bracis-33583 70 13 br br NOUN bracis-33583 70 14 could could AUX bracis-33583 70 15 not not PART bracis-33583 70 16 do do VERB bracis-33583 70 17 . . PUNCT bracis-33583 71 1 however however ADV bracis-33583 71 2 , , PUNCT bracis-33583 71 3 this this DET bracis-33583 71 4 method method NOUN bracis-33583 71 5 can can AUX bracis-33583 71 6 escalate escalate VERB bracis-33583 71 7 quickly quickly ADV bracis-33583 71 8 and and CCONJ bracis-33583 71 9 create create VERB bracis-33583 71 10 many many ADJ bracis-33583 71 11 classes class NOUN bracis-33583 71 12 with with ADP bracis-33583 71 13 few few ADJ bracis-33583 71 14 positive positive ADJ bracis-33583 71 15 instances instance NOUN bracis-33583 71 16 . . PUNCT bracis-33583 72 1 a a DET bracis-33583 72 2 modified modify VERB bracis-33583 72 3 version version NOUN bracis-33583 72 4 called call VERB bracis-33583 72 5 random random ADJ bracis-33583 72 6 k k NOUN bracis-33583 72 7 - - PUNCT bracis-33583 72 8 labelsets labelset NOUN bracis-33583 72 9 ( ( PUNCT bracis-33583 72 10 rakel rakel NOUN bracis-33583 72 11 ) ) PUNCT bracis-33583 72 12     SPACE bracis-33583 73 1 [ [ X bracis-33583 73 2 37 37 NUM bracis-33583 73 3 ] ] PUNCT bracis-33583 73 4 creates create VERB bracis-33583 73 5 combinations combination NOUN bracis-33583 73 6 of of ADP bracis-33583 73 7 label label NOUN bracis-33583 73 8 - - PUNCT bracis-33583 73 9 powerset powerset NOUN bracis-33583 73 10 classifiers classifier NOUN bracis-33583 73 11 iteratively iteratively ADV bracis-33583 73 12 , , PUNCT bracis-33583 73 13 which which PRON bracis-33583 73 14 considers consider VERB bracis-33583 73 15 label label NOUN bracis-33583 73 16 correlations correlation NOUN bracis-33583 73 17 without without ADP bracis-33583 73 18 escalating escalate VERB bracis-33583 73 19 the the DET bracis-33583 73 20 problem problem NOUN bracis-33583 73 21 . . PUNCT bracis-33583 74 1 pruned prune VERB bracis-33583 74 2 sets set NOUN bracis-33583 74 3 ( ( PUNCT bracis-33583 74 4 ps ps NOUN bracis-33583 74 5 ) ) PUNCT bracis-33583 74 6     SPACE bracis-33583 75 1 [ [ X bracis-33583 75 2 26 26 NUM bracis-33583 75 3 ] ] PUNCT bracis-33583 75 4 is be AUX bracis-33583 75 5 a a DET bracis-33583 75 6 strategy strategy NOUN bracis-33583 75 7 that that PRON bracis-33583 75 8 identifies identify VERB bracis-33583 75 9 the the DET bracis-33583 75 10 relation relation NOUN bracis-33583 75 11 of of ADP bracis-33583 75 12 the the DET bracis-33583 75 13 labels label NOUN bracis-33583 75 14 and and CCONJ bracis-33583 75 15 eliminates eliminate VERB bracis-33583 75 16 the the DET bracis-33583 75 17 ones one NOUN bracis-33583 75 18 with with ADP bracis-33583 75 19 less less ADJ bracis-33583 75 20 usage usage NOUN bracis-33583 75 21 , , PUNCT bracis-33583 75 22 grouping group VERB bracis-33583 75 23 them they PRON bracis-33583 75 24 based base VERB bracis-33583 75 25 on on ADP bracis-33583 75 26 their their PRON bracis-33583 75 27 density density NOUN bracis-33583 75 28 and and CCONJ bracis-33583 75 29 frequency frequency NOUN bracis-33583 75 30 while while SCONJ bracis-33583 75 31 trying try VERB bracis-33583 75 32 not not PART bracis-33583 75 33 to to PART bracis-33583 75 34 lose lose VERB bracis-33583 75 35 data datum NOUN bracis-33583 75 36 . . PUNCT bracis-33583 76 1 the the DET bracis-33583 76 2 ensemble ensemble ADJ bracis-33583 76 3 of of ADP bracis-33583 76 4 pruned prune VERB bracis-33583 76 5 sets set NOUN bracis-33583 76 6 ( ( PUNCT bracis-33583 76 7 eps eps NOUN bracis-33583 76 8 ) ) PUNCT bracis-33583 76 9 is be AUX bracis-33583 76 10 also also ADV bracis-33583 76 11 presented present VERB bracis-33583 76 12 as as ADP bracis-33583 76 13 a a DET bracis-33583 76 14 version version NOUN bracis-33583 76 15 of of ADP bracis-33583 76 16 the the DET bracis-33583 76 17 algorithm algorithm NOUN bracis-33583 76 18 that that PRON bracis-33583 76 19 avoids avoid VERB bracis-33583 76 20 the the DET bracis-33583 76 21 model model NOUN bracis-33583 76 22 overfitting overfitting NOUN bracis-33583 76 23 . . PUNCT bracis-33583 77 1 2.2 2.2 NUM bracis-33583 77 2 global global ADJ bracis-33583 77 3 approach approach NOUN bracis-33583 77 4 the the DET bracis-33583 77 5 global global ADJ bracis-33583 77 6 approach approach NOUN bracis-33583 77 7 creates create VERB bracis-33583 77 8 classifiers classifier NOUN bracis-33583 77 9 that that PRON bracis-33583 77 10 take take VERB bracis-33583 77 11 into into ADP bracis-33583 77 12 account account NOUN bracis-33583 77 13 label label NOUN bracis-33583 77 14 dependency dependency NOUN bracis-33583 77 15 in in ADP bracis-33583 77 16 a a DET bracis-33583 77 17 direct direct ADJ bracis-33583 77 18 way way NOUN bracis-33583 77 19 , , PUNCT bracis-33583 77 20 dealing deal VERB bracis-33583 77 21 with with ADP bracis-33583 77 22 all all DET bracis-33583 77 23 labels label NOUN bracis-33583 77 24 simultaneously simultaneously ADV bracis-33583 77 25 . . PUNCT bracis-33583 78 1 it it PRON bracis-33583 78 2 uses use VERB bracis-33583 78 3 only only ADV bracis-33583 78 4 one one NUM bracis-33583 78 5 classifier classifier NOUN bracis-33583 78 6 to to PART bracis-33583 78 7 predict predict VERB bracis-33583 78 8 the the DET bracis-33583 78 9 l l NOUN bracis-33583 78 10 labels label NOUN bracis-33583 78 11 of of ADP bracis-33583 78 12 the the DET bracis-33583 78 13 problem problem NOUN bracis-33583 78 14 . . PUNCT bracis-33583 79 1 the the DET bracis-33583 79 2 advantage advantage NOUN bracis-33583 79 3 is be AUX bracis-33583 79 4 that that SCONJ bracis-33583 79 5 it it PRON bracis-33583 79 6 considers consider VERB bracis-33583 79 7 the the DET bracis-33583 79 8 relationships relationship NOUN bracis-33583 79 9 between between ADP bracis-33583 79 10 labels label NOUN bracis-33583 79 11 and and CCONJ bracis-33583 79 12 is be AUX bracis-33583 79 13 usually usually ADV bracis-33583 79 14 faster fast ADJ bracis-33583 79 15 to to PART bracis-33583 79 16 execute execute VERB bracis-33583 79 17 . . PUNCT bracis-33583 80 1 however however ADV bracis-33583 80 2 , , PUNCT bracis-33583 80 3 the the DET bracis-33583 80 4 approach approach NOUN bracis-33583 80 5 does do AUX bracis-33583 80 6 not not PART bracis-33583 80 7 give give VERB bracis-33583 80 8 too too ADV bracis-33583 80 9 much much ADJ bracis-33583 80 10 weight weight NOUN bracis-33583 80 11 to to ADP bracis-33583 80 12 each each DET bracis-33583 80 13 individual individual ADJ bracis-33583 80 14 label label NOUN bracis-33583 80 15 ’s ’s PART bracis-33583 80 16 peculiarity peculiarity NOUN bracis-33583 80 17 . . PUNCT bracis-33583 81 1 additionally additionally ADV bracis-33583 81 2 , , PUNCT bracis-33583 81 3 the the DET bracis-33583 81 4 algorithm algorithm NOUN bracis-33583 81 5 is be AUX bracis-33583 81 6 more more ADV bracis-33583 81 7 difficult difficult ADJ bracis-33583 81 8 to to PART bracis-33583 81 9 parallelize parallelize VERB bracis-33583 81 10 , , PUNCT bracis-33583 81 11 bringing bring VERB bracis-33583 81 12 more more ADJ bracis-33583 81 13 computational computational ADJ bracis-33583 81 14 cost cost NOUN bracis-33583 81 15 to to ADP bracis-33583 81 16 the the DET bracis-33583 81 17 algorithms algorithm NOUN bracis-33583 81 18 and and CCONJ bracis-33583 81 19 spending spend VERB bracis-33583 81 20 more more ADJ bracis-33583 81 21 training training NOUN bracis-33583 81 22 time time NOUN bracis-33583 81 23 . . PUNCT bracis-33583 82 1 one one NUM bracis-33583 82 2 of of ADP bracis-33583 82 3 the the DET bracis-33583 82 4 first first ADJ bracis-33583 82 5 global global ADV bracis-33583 82 6 - - PUNCT bracis-33583 82 7 based base VERB bracis-33583 82 8 strategies strategy NOUN bracis-33583 82 9 is be AUX bracis-33583 82 10 c4.5 c4.5 PROPN bracis-33583 82 11 m m NOUN bracis-33583 82 12     SPACE bracis-33583 83 1 [ [ X bracis-33583 83 2 16 16 NUM bracis-33583 83 3 ] ] PUNCT bracis-33583 83 4 , , PUNCT bracis-33583 83 5 a a DET bracis-33583 83 6 decision decision NOUN bracis-33583 83 7 - - PUNCT bracis-33583 83 8 tree tree NOUN bracis-33583 83 9 - - PUNCT bracis-33583 83 10 based base VERB bracis-33583 83 11 method method NOUN bracis-33583 83 12 created create VERB bracis-33583 83 13 by by ADP bracis-33583 83 14 adapting adapt VERB bracis-33583 83 15 the the DET bracis-33583 83 16 c4.5 c4.5 PROPN bracis-33583 83 17 algorithm algorithm NOUN bracis-33583 83 18 to to ADP bracis-33583 83 19 multi multi ADJ bracis-33583 83 20 - - ADJ bracis-33583 83 21 label label ADJ bracis-33583 83 22 problems problem NOUN bracis-33583 83 23 . . PUNCT bracis-33583 84 1 this this PRON bracis-33583 84 2 was be AUX bracis-33583 84 3 done do VERB bracis-33583 84 4 by by ADP bracis-33583 84 5 altering alter VERB bracis-33583 84 6 the the DET bracis-33583 84 7 entropy entropy NOUN bracis-33583 84 8 formula formula NOUN bracis-33583 84 9 . . PUNCT bracis-33583 85 1 the the DET bracis-33583 85 2 algorithm algorithm NOUN bracis-33583 85 3 was be AUX bracis-33583 85 4 also also ADV bracis-33583 85 5 changed change VERB bracis-33583 85 6 to to PART bracis-33583 85 7 represent represent VERB bracis-33583 85 8 a a DET bracis-33583 85 9 set set NOUN bracis-33583 85 10 of of ADP bracis-33583 85 11 labels label NOUN bracis-33583 85 12 in in ADP bracis-33583 85 13 their their PRON bracis-33583 85 14 tree tree NOUN bracis-33583 85 15 leaf leaf NOUN bracis-33583 85 16 nodes node NOUN bracis-33583 85 17 . . PUNCT bracis-33583 86 1 the the DET bracis-33583 86 2 mlknn mlknn ADJ bracis-33583 86 3 method method NOUN bracis-33583 86 4 [ [ X bracis-33583 86 5 41 41 NUM bracis-33583 86 6 , , PUNCT bracis-33583 86 7 43 43 NUM bracis-33583 86 8 ] ] PUNCT bracis-33583 86 9 is be AUX bracis-33583 86 10 based base VERB bracis-33583 86 11 on on ADP bracis-33583 86 12 the the DET bracis-33583 86 13 knn knn PROPN bracis-33583 86 14 algorithm algorithm PROPN bracis-33583 86 15 where where SCONJ bracis-33583 86 16 , , PUNCT bracis-33583 86 17 for for ADP bracis-33583 86 18 each each DET bracis-33583 86 19 instance instance NOUN bracis-33583 86 20 , , PUNCT bracis-33583 86 21 the the DET bracis-33583 86 22 k k PROPN bracis-33583 86 23 closest close ADJ bracis-33583 86 24 neighbors neighbor NOUN bracis-33583 86 25 classes class NOUN bracis-33583 86 26 are be AUX bracis-33583 86 27 compared compare VERB bracis-33583 86 28 . . PUNCT bracis-33583 87 1 based base VERB bracis-33583 87 2 on on ADP bracis-33583 87 3 this this PRON bracis-33583 87 4 , , PUNCT bracis-33583 87 5 the the DET bracis-33583 87 6 instance instance NOUN bracis-33583 87 7 is be AUX bracis-33583 87 8 labeled label VERB bracis-33583 87 9 . . PUNCT bracis-33583 88 1 in in ADP bracis-33583 88 2 zhang zhang PROPN bracis-33583 88 3 and and CCONJ bracis-33583 88 4 zhou zhou PROPN bracis-33583 88 5     SPACE bracis-33583 89 1 [ [ X bracis-33583 89 2 42 42 NUM bracis-33583 89 3 ] ] PUNCT bracis-33583 89 4 , , PUNCT bracis-33583 89 5 the the DET bracis-33583 89 6 back back ADJ bracis-33583 89 7 - - PUNCT bracis-33583 89 8 propagation propagation NOUN bracis-33583 89 9 multi multi ADJ bracis-33583 89 10 - - ADJ bracis-33583 89 11 label label ADJ bracis-33583 89 12 learning learning NOUN bracis-33583 89 13 was be AUX bracis-33583 89 14 proposed propose VERB bracis-33583 89 15 , , PUNCT bracis-33583 89 16 in in ADP bracis-33583 89 17 which which PRON bracis-33583 89 18 a a DET bracis-33583 89 19 multi multi ADJ bracis-33583 89 20 - - ADJ bracis-33583 89 21 label label ADJ bracis-33583 89 22 error error NOUN bracis-33583 89 23 measure measure NOUN bracis-33583 89 24 is be AUX bracis-33583 89 25 proposed propose VERB bracis-33583 89 26 and and CCONJ bracis-33583 89 27 used use VERB bracis-33583 89 28 in in ADP bracis-33583 89 29 the the DET bracis-33583 89 30 output output NOUN bracis-33583 89 31 layer layer NOUN bracis-33583 89 32 of of ADP bracis-33583 89 33 an an DET bracis-33583 89 34 artificial artificial ADJ bracis-33583 89 35 neural neural ADJ bracis-33583 89 36 network network NOUN bracis-33583 89 37 . . PUNCT bracis-33583 90 1 in in ADP bracis-33583 90 2 tsoumakas tsoumakas PROPN bracis-33583 90 3 et et PROPN bracis-33583 90 4 al al PROPN bracis-33583 90 5 . . PUNCT bracis-33583 90 6     SPACE bracis-33583 91 1 [ [ X bracis-33583 91 2 34 34 NUM bracis-33583 91 3 ] ] PUNCT bracis-33583 91 4 , , PUNCT bracis-33583 91 5 the the DET bracis-33583 91 6 hierarchy hierarchy NOUN bracis-33583 91 7 of of ADP bracis-33583 91 8 multi multi ADJ bracis-33583 91 9 - - ADJ bracis-33583 91 10 label label ADJ bracis-33583 91 11 classifiers classifier NOUN bracis-33583 91 12 ( ( PUNCT bracis-33583 91 13 homer homer NOUN bracis-33583 91 14 ) ) PUNCT bracis-33583 91 15 is be AUX bracis-33583 91 16 proposed propose VERB bracis-33583 91 17 to to PART bracis-33583 91 18 deal deal VERB bracis-33583 91 19 with with ADP bracis-33583 91 20 large large ADJ bracis-33583 91 21 label label NOUN bracis-33583 91 22 sets set NOUN bracis-33583 91 23 , , PUNCT bracis-33583 91 24 focusing focus VERB bracis-33583 91 25 on on ADP bracis-33583 91 26 being be AUX bracis-33583 91 27 effective effective ADJ bracis-33583 91 28 and and CCONJ bracis-33583 91 29 computationally computationally ADV bracis-33583 91 30 efficient efficient ADJ bracis-33583 91 31 . . PUNCT bracis-33583 92 1 this this DET bracis-33583 92 2 method method NOUN bracis-33583 92 3 uses use VERB bracis-33583 92 4 a a DET bracis-33583 92 5 “ " PUNCT bracis-33583 92 6 divide divide NOUN bracis-33583 92 7 and and CCONJ bracis-33583 92 8 conquer conquer VERB bracis-33583 92 9 ” " PUNCT bracis-33583 92 10 label label NOUN bracis-33583 92 11 hierarchy hierarchy NOUN bracis-33583 92 12 strategy strategy NOUN bracis-33583 92 13 . . PUNCT bracis-33583 93 1 it it PRON bracis-33583 93 2 was be AUX bracis-33583 93 3 able able ADJ bracis-33583 93 4 to to PART bracis-33583 93 5 have have VERB bracis-33583 93 6 linear linear ADJ bracis-33583 93 7 training training NOUN bracis-33583 93 8 and and CCONJ bracis-33583 93 9 logarithmic logarithmic ADJ bracis-33583 93 10 testing testing NOUN bracis-33583 93 11 complexity complexity NOUN bracis-33583 93 12 . . PUNCT bracis-33583 94 1 colombini colombini PROPN bracis-33583 94 2 et et PROPN bracis-33583 94 3 al al PROPN bracis-33583 94 4 . . PUNCT bracis-33583 94 5     SPACE bracis-33583 95 1 [ [ X bracis-33583 95 2 17 17 NUM bracis-33583 95 3 ] ] PUNCT bracis-33583 95 4 proposed propose VERB bracis-33583 95 5 som som PROPN bracis-33583 95 6 - - PUNCT bracis-33583 95 7 mll mll PROPN bracis-33583 95 8 ( ( PUNCT bracis-33583 95 9 self self NOUN bracis-33583 95 10 - - PUNCT bracis-33583 95 11 organizing organize VERB bracis-33583 95 12 maps map NOUN bracis-33583 95 13 multi multi ADJ bracis-33583 95 14 - - ADJ bracis-33583 95 15 label label ADJ bracis-33583 95 16 learning learning NOUN bracis-33583 95 17 ) ) PUNCT bracis-33583 95 18 , , PUNCT bracis-33583 95 19 implementing implement VERB bracis-33583 95 20 kohonen kohonen PROPN bracis-33583 95 21 maps map NOUN bracis-33583 95 22 to to PART bracis-33583 95 23 predict predict VERB bracis-33583 95 24 multi multi ADJ bracis-33583 95 25 - - ADJ bracis-33583 95 26 label label ADJ bracis-33583 95 27 data datum NOUN bracis-33583 95 28 . . PUNCT bracis-33583 96 1 this this DET bracis-33583 96 2 algorithm algorithm NOUN bracis-33583 96 3 was be AUX bracis-33583 96 4 also also ADV bracis-33583 96 5 used use VERB bracis-33583 96 6 by by ADP bracis-33583 96 7 alshanqiti alshanqiti ADV bracis-33583 96 8 and and CCONJ bracis-33583 96 9 namoun namoun NOUN bracis-33583 96 10     SPACE bracis-33583 97 1 [ [ X bracis-33583 97 2 7 7 X bracis-33583 97 3 ] ] PUNCT bracis-33583 97 4 to to PART bracis-33583 97 5 compare compare VERB bracis-33583 97 6 the the DET bracis-33583 97 7 method method NOUN bracis-33583 97 8 with with ADP bracis-33583 97 9 other other ADJ bracis-33583 97 10 baseline baseline NOUN bracis-33583 97 11 models model NOUN bracis-33583 97 12 , , PUNCT bracis-33583 97 13 like like ADP bracis-33583 97 14 linear linear PROPN bracis-33583 97 15 regression regression NOUN bracis-33583 97 16 and and CCONJ bracis-33583 97 17 matrix matrix NOUN bracis-33583 97 18 factorization factorization NOUN bracis-33583 97 19 . . PUNCT bracis-33583 98 1 in in ADP bracis-33583 98 2 saini saini PROPN bracis-33583 98 3 et et PROPN bracis-33583 98 4 al al PROPN bracis-33583 98 5 . . PUNCT bracis-33583 98 6     SPACE bracis-33583 99 1 [ [ X bracis-33583 99 2 28 28 NUM bracis-33583 99 3 ] ] X bracis-33583 99 4 , , PUNCT bracis-33583 99 5 a a DET bracis-33583 99 6 different different ADJ bracis-33583 99 7 structure structure NOUN bracis-33583 99 8 of of ADP bracis-33583 99 9 som som NOUN bracis-33583 99 10 for for ADP bracis-33583 99 11 multi multi ADJ bracis-33583 99 12 - - ADJ bracis-33583 99 13 label label ADJ bracis-33583 99 14 data datum NOUN bracis-33583 99 15 was be AUX bracis-33583 99 16 presented present VERB bracis-33583 99 17 and and CCONJ bracis-33583 99 18 called call VERB bracis-33583 99 19 ml ml NOUN bracis-33583 99 20 - - PUNCT bracis-33583 99 21 som som NOUN bracis-33583 99 22 , , PUNCT bracis-33583 99 23 where where SCONJ bracis-33583 99 24 the the DET bracis-33583 99 25 label label NOUN bracis-33583 99 26 vector vector NOUN bracis-33583 99 27 is be AUX bracis-33583 99 28 determined determine VERB bracis-33583 99 29 using use VERB bracis-33583 99 30 the the DET bracis-33583 99 31 closest close ADJ bracis-33583 99 32 and and CCONJ bracis-33583 99 33 neighboring neighboring NOUN bracis-33583 99 34 neurons neuron NOUN bracis-33583 99 35 , , PUNCT bracis-33583 99 36 obtaining obtain VERB bracis-33583 99 37 better well ADJ bracis-33583 99 38 results result NOUN bracis-33583 99 39 compared compare VERB bracis-33583 99 40 to to ADP bracis-33583 99 41 som som NOUN bracis-33583 99 42 - - PUNCT bracis-33583 99 43 mll mll PROPN bracis-33583 99 44 in in ADP bracis-33583 99 45 some some DET bracis-33583 99 46     SPACE bracis-33583 99 47 cases case NOUN bracis-33583 99 48 . . PUNCT bracis-33583 100 1 3 3 NUM bracis-33583 100 2 multi multi ADJ bracis-33583 100 3 - - ADJ bracis-33583 100 4 label label ADJ bracis-33583 100 5 classification classification NOUN bracis-33583 100 6 with with ADP bracis-33583 100 7     SPACE bracis-33583 100 8 self self NOUN bracis-33583 100 9 - - PUNCT bracis-33583 100 10 organizing organize VERB bracis-33583 100 11 maps map NOUN bracis-33583 100 12 this this DET bracis-33583 100 13 section section NOUN bracis-33583 100 14 provides provide VERB bracis-33583 100 15 an an DET bracis-33583 100 16 overview overview NOUN bracis-33583 100 17 of of ADP bracis-33583 100 18 self self NOUN bracis-33583 100 19 - - PUNCT bracis-33583 100 20 organizing organize VERB bracis-33583 100 21 maps map NOUN bracis-33583 100 22 ( ( PUNCT bracis-33583 100 23 kohonen kohonen PROPN bracis-33583 100 24 maps maps PROPN bracis-33583 100 25 ) ) PUNCT bracis-33583 100 26 and and CCONJ bracis-33583 100 27 the the DET bracis-33583 100 28 self self NOUN bracis-33583 100 29 - - PUNCT bracis-33583 100 30 organizing organize VERB bracis-33583 100 31 map map NOUN bracis-33583 100 32 multi multi ADJ bracis-33583 100 33 - - ADJ bracis-33583 100 34 label label ADJ bracis-33583 100 35 learning learning NOUN bracis-33583 100 36 ( ( PUNCT bracis-33583 100 37 som som NOUN bracis-33583 100 38 - - PUNCT bracis-33583 100 39 mll mll PROPN bracis-33583 100 40 ) ) PUNCT bracis-33583 100 41     SPACE bracis-33583 101 1 [ [ X bracis-33583 101 2 17 17 NUM bracis-33583 101 3 ] ] SYM bracis-33583 101 4 . . PUNCT bracis-33583 102 1 3.1 3.1 NUM bracis-33583 102 2 self self NOUN bracis-33583 102 3 - - PUNCT bracis-33583 102 4 organizing organize VERB bracis-33583 102 5 maps map NOUN bracis-33583 102 6 self self NOUN bracis-33583 102 7 - - PUNCT bracis-33583 102 8 organizing organize VERB bracis-33583 102 9 maps map NOUN bracis-33583 102 10 ( ( PUNCT bracis-33583 102 11 som som NOUN bracis-33583 102 12 ) ) PUNCT bracis-33583 102 13     SPACE bracis-33583 103 1 [ [ X bracis-33583 103 2 21 21 NUM bracis-33583 103 3 ] ] PUNCT bracis-33583 103 4 are be AUX bracis-33583 103 5 unsupervised unsupervised ADJ bracis-33583 103 6 neural neural ADJ bracis-33583 103 7 networks network NOUN bracis-33583 103 8 with with ADP bracis-33583 103 9 a a DET bracis-33583 103 10 fixed fix VERB bracis-33583 103 11 - - PUNCT bracis-33583 103 12 size size NOUN bracis-33583 103 13 bi bi ADJ bracis-33583 103 14 - - ADJ bracis-33583 103 15 dimensional dimensional ADJ bracis-33583 103 16 grid grid NOUN bracis-33583 103 17 of of ADP bracis-33583 103 18 neurons neuron NOUN bracis-33583 103 19 . . PUNCT bracis-33583 104 1 when when SCONJ bracis-33583 104 2 an an DET bracis-33583 104 3 instance instance NOUN bracis-33583 104 4 is be AUX bracis-33583 104 5 mapped map VERB bracis-33583 104 6 to to ADP bracis-33583 104 7 the the DET bracis-33583 104 8 grid grid NOUN bracis-33583 104 9 , , PUNCT bracis-33583 104 10 its its PRON bracis-33583 104 11 best well ADV bracis-33583 104 12 - - PUNCT bracis-33583 104 13 matching match VERB bracis-33583 104 14 neuron neuron NOUN bracis-33583 104 15 is be AUX bracis-33583 104 16 selected select VERB bracis-33583 104 17 as as ADP bracis-33583 104 18 the the DET bracis-33583 104 19 winning winning NOUN bracis-33583 104 20 neuron neuron NOUN bracis-33583 104 21 , , PUNCT bracis-33583 104 22 which which PRON bracis-33583 104 23 strengthens strengthen VERB bracis-33583 104 24 its its PRON bracis-33583 104 25 proximity proximity NOUN bracis-33583 104 26 to to ADP bracis-33583 104 27 the the DET bracis-33583 104 28 instance instance NOUN bracis-33583 104 29 . . PUNCT bracis-33583 105 1 the the DET bracis-33583 105 2 mapping mapping NOUN bracis-33583 105 3 process process NOUN bracis-33583 105 4 is be AUX bracis-33583 105 5 illustrated illustrate VERB bracis-33583 105 6 in in ADP bracis-33583 105 7 fig fig NOUN bracis-33583 105 8 . . PUNCT bracis-33583 105 9     SPACE bracis-33583 106 1 1 1 NUM bracis-33583 106 2 . . PUNCT bracis-33583 106 3 fig fig NOUN bracis-33583 106 4 . . PUNCT bracis-33583 107 1 1 1 X bracis-33583 107 2 . . X bracis-33583 107 3 mapping map VERB bracis-33583 107 4 an an DET bracis-33583 107 5 input input NOUN bracis-33583 107 6 instance instance NOUN bracis-33583 107 7 in in ADP bracis-33583 107 8 the the DET bracis-33583 107 9 kohonen kohonen PROPN bracis-33583 107 10 map map NOUN bracis-33583 107 11 full full ADJ bracis-33583 107 12 size size NOUN bracis-33583 107 13 image image NOUN bracis-33583 107 14 the the DET bracis-33583 107 15 training training NOUN bracis-33583 107 16 starts start VERB bracis-33583 107 17 by by ADP bracis-33583 107 18 calculating calculate VERB bracis-33583 107 19 the the DET bracis-33583 107 20 euclidean euclidean ADJ bracis-33583 107 21 distance distance NOUN bracis-33583 107 22 between between ADP bracis-33583 107 23 the the DET bracis-33583 107 24 input input NOUN bracis-33583 107 25 instance instance NOUN bracis-33583 107 26 and and CCONJ bracis-33583 107 27 all all DET bracis-33583 107 28 neuron neuron NOUN bracis-33583 107 29 weight weight NOUN bracis-33583 107 30 vectors vector NOUN bracis-33583 107 31 . . PUNCT bracis-33583 108 1 the the DET bracis-33583 108 2 winning win VERB bracis-33583 108 3 neuron neuron NOUN bracis-33583 108 4 is be AUX bracis-33583 108 5 the the DET bracis-33583 108 6 one one NUM bracis-33583 108 7 with with ADP bracis-33583 108 8 the the DET bracis-33583 108 9 minimum minimum ADJ bracis-33583 108 10 distance distance NOUN bracis-33583 108 11 according accord VERB bracis-33583 108 12 to to ADP bracis-33583 108 13 eq eq PROPN bracis-33583 108 14 . . PUNCT bracis-33583 108 15     SPACE bracis-33583 108 16 1 1 NUM bracis-33583 108 17 . . PUNCT bracis-33583 109 1 in in ADP bracis-33583 109 2 this this DET bracis-33583 109 3 equation equation NOUN bracis-33583 109 4 , , PUNCT bracis-33583 109 5 \(d_j\ \(d_j\ PROPN bracis-33583 109 6 ) ) PUNCT bracis-33583 109 7 is be AUX bracis-33583 109 8 the the DET bracis-33583 109 9 distance distance NOUN bracis-33583 109 10 between between ADP bracis-33583 109 11 the the DET bracis-33583 109 12 input input NOUN bracis-33583 109 13 data data PROPN bracis-33583 109 14 \(\textbf{x}\ \(\textbf{x}\ NOUN bracis-33583 109 15 ) ) PUNCT bracis-33583 109 16 and and CCONJ bracis-33583 109 17 the the DET bracis-33583 109 18 neuron neuron PROPN bracis-33583 109 19 j j PROPN bracis-33583 109 20 , , PUNCT bracis-33583 109 21 with with ADP bracis-33583 109 22 a a DET bracis-33583 109 23 the the DET bracis-33583 109 24 dimension dimension NOUN bracis-33583 109 25 of of ADP bracis-33583 109 26 the the DET bracis-33583 109 27 input input NOUN bracis-33583 109 28 instance instance NOUN bracis-33583 109 29 . . PUNCT bracis-33583 110 1 each each DET bracis-33583 110 2 neuron neuron NOUN bracis-33583 110 3 weight weight NOUN bracis-33583 110 4 vector vector NOUN bracis-33583 110 5 is be AUX bracis-33583 110 6 initialized initialize VERB bracis-33583 110 7 with with ADP bracis-33583 110 8 small small ADJ bracis-33583 110 9 random random ADJ bracis-33583 110 10 values value NOUN bracis-33583 110 11 . . PUNCT bracis-33583 111 1 $ $ SYM bracis-33583 111 2 $ $ SYM bracis-33583 111 3 \begin{aligned \begin{aligne VERB bracis-33583 111 4 } } PUNCT bracis-33583 111 5 d_j d_j PUNCT bracis-33583 111 6 ( ( PUNCT bracis-33583 111 7 x x X bracis-33583 111 8 ) ) PUNCT bracis-33583 111 9 = = SYM bracis-33583 111 10 \sqrt{\sum \sqrt{\sum NOUN bracis-33583 111 11 _ _ PUNCT bracis-33583 111 12 { { PUNCT bracis-33583 111 13 i=1}^{a i=1}^{a NOUN bracis-33583 111 14 } } PUNCT bracis-33583 111 15 { { PUNCT bracis-33583 111 16 ( ( PUNCT bracis-33583 111 17 x_i x_i PUNCT bracis-33583 111 18 w_{j w_{j NOUN bracis-33583 111 19 , , PUNCT bracis-33583 111 20 i})^2 i})^2 PROPN bracis-33583 111 21 } } PUNCT bracis-33583 111 22 } } PUNCT bracis-33583 111 23 \end{aligned}$$ \end{aligned}$$ X bracis-33583 111 24 ( ( PUNCT bracis-33583 111 25 1 1 NUM bracis-33583 111 26 ) ) PUNCT bracis-33583 111 27 as as SCONJ bracis-33583 111 28 the the DET bracis-33583 111 29 winner winner NOUN bracis-33583 111 30 neuron neuron NOUN bracis-33583 111 31 is be AUX bracis-33583 111 32 defined define VERB bracis-33583 111 33 , , PUNCT bracis-33583 111 34 the the DET bracis-33583 111 35 value value NOUN bracis-33583 111 36 of of ADP bracis-33583 111 37 its its PRON bracis-33583 111 38 weights weight NOUN bracis-33583 111 39 is be AUX bracis-33583 111 40 updated update VERB bracis-33583 111 41 based base VERB bracis-33583 111 42 on on ADP bracis-33583 111 43 the the DET bracis-33583 111 44 input input NOUN bracis-33583 111 45 data datum NOUN bracis-33583 111 46 and and CCONJ bracis-33583 111 47 the the DET bracis-33583 111 48 weights weight NOUN bracis-33583 111 49 of of ADP bracis-33583 111 50 the the DET bracis-33583 111 51 neurons neuron NOUN bracis-33583 111 52 in in ADP bracis-33583 111 53 its its PRON bracis-33583 111 54 neighborhood neighborhood NOUN bracis-33583 111 55 , , PUNCT bracis-33583 111 56 according accord VERB bracis-33583 111 57 to to ADP bracis-33583 111 58 eq eq PROPN bracis-33583 111 59 . . PUNCT bracis-33583 111 60     SPACE bracis-33583 111 61 2 2 X bracis-33583 111 62 . . PUNCT bracis-33583 112 1 the the DET bracis-33583 112 2 hyperparameter hyperparameter NOUN bracis-33583 112 3 \(\alpha \(\alpha ADP bracis-33583 112 4 \ \ NOUN bracis-33583 112 5 ) ) PUNCT bracis-33583 112 6 is be AUX bracis-33583 112 7 the the DET bracis-33583 112 8 learning learn VERB bracis-33583 112 9 rate rate NOUN bracis-33583 112 10 of of ADP bracis-33583 112 11 the the DET bracis-33583 112 12 model model NOUN bracis-33583 112 13 , , PUNCT bracis-33583 112 14 which which PRON bracis-33583 112 15 typically typically ADV bracis-33583 112 16 decreases decrease VERB bracis-33583 112 17 with with ADP bracis-33583 112 18 training training NOUN bracis-33583 112 19 time time NOUN bracis-33583 112 20 . . PUNCT bracis-33583 113 1 $ $ SYM bracis-33583 113 2 $ $ SYM bracis-33583 113 3 \begin{aligned \begin{aligne VERB bracis-33583 113 4 } } PUNCT bracis-33583 113 5 w_j w_j NUM bracis-33583 113 6 ( ( PUNCT bracis-33583 113 7 t+1 t+1 NOUN bracis-33583 113 8 ) ) PUNCT bracis-33583 113 9 = = PRON bracis-33583 113 10 w_j(t w_j(t PROPN bracis-33583 113 11 ) ) PUNCT bracis-33583 114 1 + + CCONJ bracis-33583 114 2 \alpha \alpha PROPN bracis-33583 114 3 ( ( PUNCT bracis-33583 114 4 t t PROPN bracis-33583 114 5 ) ) PUNCT bracis-33583 114 6 \times \times ADP bracis-33583 114 7 h_{j h_{j PROPN bracis-33583 114 8 , , PUNCT bracis-33583 114 9 i i NOUN bracis-33583 114 10 } } PUNCT bracis-33583 114 11 \times \times PROPN bracis-33583 114 12 ( ( PUNCT bracis-33583 114 13 \textbf{x \textbf{x NOUN bracis-33583 114 14 } } PUNCT bracis-33583 114 15 w_j(t w_j(t PROPN bracis-33583 114 16 ) ) PUNCT bracis-33583 114 17 ) ) PUNCT bracis-33583 114 18 \end{aligned}$$ \end{aligned}$$ PROPN bracis-33583 114 19 ( ( PUNCT bracis-33583 114 20 2 2 X bracis-33583 114 21 ) ) PUNCT bracis-33583 114 22 the the DET bracis-33583 114 23 \(h_{j \(h_{j NOUN bracis-33583 114 24 , , PUNCT bracis-33583 114 25 i}\ i}\ NUM bracis-33583 114 26 ) ) PUNCT bracis-33583 114 27 is be AUX bracis-33583 114 28 the the DET bracis-33583 114 29 influence influence NOUN bracis-33583 114 30 that that PRON bracis-33583 114 31 the the DET bracis-33583 114 32 winning win VERB bracis-33583 114 33 neuron neuron PROPN bracis-33583 114 34 j j PROPN bracis-33583 114 35 has have VERB bracis-33583 114 36 in in ADP bracis-33583 114 37 its its PRON bracis-33583 114 38 neighborhood neighborhood NOUN bracis-33583 114 39 of of ADP bracis-33583 114 40 neurons neuron NOUN bracis-33583 114 41 . . PUNCT bracis-33583 115 1 the the DET bracis-33583 115 2 most most ADV bracis-33583 115 3 commonly commonly ADV bracis-33583 115 4 used use VERB bracis-33583 115 5 formula formula NOUN bracis-33583 115 6 is be AUX bracis-33583 115 7 the the DET bracis-33583 115 8 gaussian gaussian NOUN bracis-33583 115 9 , , PUNCT bracis-33583 115 10 which which PRON bracis-33583 115 11 can can AUX bracis-33583 115 12 be be AUX bracis-33583 115 13 seen see VERB bracis-33583 115 14 in in ADP bracis-33583 115 15 eq eq NOUN bracis-33583 115 16 . . PUNCT bracis-33583 115 17     SPACE bracis-33583 115 18 3 3 X bracis-33583 115 19 . . PUNCT bracis-33583 116 1 the the DET bracis-33583 116 2 \(\sigma \(\sigma NOUN bracis-33583 116 3 \ \ NOUN bracis-33583 116 4 ) ) PUNCT bracis-33583 116 5 hyperparameter hyperparameter NOUN bracis-33583 116 6 defines define VERB bracis-33583 116 7 the the DET bracis-33583 116 8 influence influence NOUN bracis-33583 116 9 ’s ’s PART bracis-33583 116 10 size size NOUN bracis-33583 116 11 , , PUNCT bracis-33583 116 12 i.e. i.e. X bracis-33583 116 13 , , PUNCT bracis-33583 116 14 the the DET bracis-33583 116 15 neighborhood neighborhood NOUN bracis-33583 116 16 width width NOUN bracis-33583 116 17 . . PUNCT bracis-33583 117 1 $ $ SYM bracis-33583 117 2 $ $ SYM bracis-33583 117 3 \begin{aligned \begin{aligne VERB bracis-33583 117 4 } } PUNCT bracis-33583 117 5 h_{j h_{j NOUN bracis-33583 117 6 , , PUNCT bracis-33583 117 7 i i NOUN bracis-33583 117 8 } } PUNCT bracis-33583 117 9 = = PUNCT bracis-33583 118 1 \exp \exp PROPN bracis-33583 118 2 \biggl \biggl PROPN bracis-33583 118 3 ( ( PUNCT bracis-33583 118 4 \ \ PROPN bracis-33583 118 5 , , PUNCT bracis-33583 118 6 -\frac{d_{j}^2(x)}{2\sigma -\frac{d_{j}^2(x)}{2\sigma PUNCT bracis-33583 118 7 ^2 ^2 ADP bracis-33583 118 8 } } PUNCT bracis-33583 118 9 \biggl \biggl PROPN bracis-33583 118 10 ) ) PUNCT bracis-33583 118 11 \end{aligned}$$ \end{aligned}$$ PROPN bracis-33583 118 12 ( ( PUNCT bracis-33583 118 13 3 3 X bracis-33583 118 14 ) ) PUNCT bracis-33583 118 15 the the DET bracis-33583 118 16 self self NOUN bracis-33583 118 17 - - PUNCT bracis-33583 118 18 organizing organize VERB bracis-33583 118 19 map map NOUN bracis-33583 118 20 multi multi ADJ bracis-33583 118 21 - - ADJ bracis-33583 118 22 label label ADJ bracis-33583 118 23 learning learning NOUN bracis-33583 118 24 ( ( PUNCT bracis-33583 118 25 som som NOUN bracis-33583 118 26 - - PUNCT bracis-33583 118 27 mll mll PROPN bracis-33583 118 28 ) ) PUNCT bracis-33583 118 29 is be AUX bracis-33583 118 30 trained train VERB bracis-33583 118 31 similarly similarly ADV bracis-33583 118 32 to to ADP bracis-33583 118 33 the the DET bracis-33583 118 34 common common ADJ bracis-33583 118 35 som som NOUN bracis-33583 118 36 . . PUNCT bracis-33583 119 1 the the DET bracis-33583 119 2 difference difference NOUN bracis-33583 119 3 is be AUX bracis-33583 119 4 that that SCONJ bracis-33583 119 5 in in ADP bracis-33583 119 6 the the DET bracis-33583 119 7 prediction prediction NOUN bracis-33583 119 8 process process NOUN bracis-33583 119 9 , , PUNCT bracis-33583 119 10 multiple multiple ADJ bracis-33583 119 11 labels label NOUN bracis-33583 119 12 are be AUX bracis-33583 119 13 considered consider VERB bracis-33583 119 14 instead instead ADV bracis-33583 119 15 of of ADP bracis-33583 119 16 only only ADV bracis-33583 119 17 one one NUM bracis-33583 119 18 label label NOUN bracis-33583 119 19 . . PUNCT bracis-33583 120 1 3.2 3.2 NUM bracis-33583 120 2 predicting predict VERB bracis-33583 120 3 new new ADJ bracis-33583 120 4 data datum NOUN bracis-33583 120 5 in in ADP bracis-33583 120 6 the the DET bracis-33583 120 7 som som PROPN bracis-33583 120 8 - - PUNCT bracis-33583 120 9 mll mll PROPN bracis-33583 120 10 , , PUNCT bracis-33583 120 11 every every DET bracis-33583 120 12 instance instance NOUN bracis-33583 120 13 \(\textbf{x}_i\ \(\textbf{x}_i\ PROPN bracis-33583 120 14 ) ) PUNCT bracis-33583 120 15 is be AUX bracis-33583 120 16 associated associate VERB bracis-33583 120 17 with with ADP bracis-33583 120 18 a a DET bracis-33583 120 19 binary binary ADJ bracis-33583 120 20 vector vector NOUN bracis-33583 120 21 \(\boldsymbol{v}_i\ \(\boldsymbol{v}_i\ NOUN bracis-33583 120 22 ) ) PUNCT bracis-33583 120 23 , , PUNCT bracis-33583 120 24 which which PRON bracis-33583 120 25 has have VERB bracis-33583 120 26 the the DET bracis-33583 120 27 value value NOUN bracis-33583 120 28 1 1 NUM bracis-33583 120 29 in in ADP bracis-33583 120 30 position position NOUN bracis-33583 120 31 j j PROPN bracis-33583 120 32 if if SCONJ bracis-33583 120 33 the the DET bracis-33583 120 34 instance instance NOUN bracis-33583 120 35 is be AUX bracis-33583 120 36 classified classify VERB bracis-33583 120 37 in in ADP bracis-33583 120 38 class class NOUN bracis-33583 120 39 \(c_j\ \(c_j\ NOUN bracis-33583 120 40 ) ) PUNCT bracis-33583 120 41 , , PUNCT bracis-33583 120 42 and and CCONJ bracis-33583 120 43 0 0 NUM bracis-33583 120 44 otherwise otherwise ADV bracis-33583 120 45 . . PUNCT bracis-33583 121 1 after after ADP bracis-33583 121 2 training training NOUN bracis-33583 121 3 , , PUNCT bracis-33583 121 4 these these DET bracis-33583 121 5 binary binary ADJ bracis-33583 121 6 vectors vector NOUN bracis-33583 121 7 are be AUX bracis-33583 121 8 used use VERB bracis-33583 121 9 to to PART bracis-33583 121 10 construct construct VERB bracis-33583 121 11 a a DET bracis-33583 121 12 prototype prototype NOUN bracis-33583 121 13 vector vector NOUN bracis-33583 121 14 \(\overline{\boldsymbol{v}}_i\ \(\overline{\boldsymbol{v}}_i\ NOUN bracis-33583 121 15 ) ) PUNCT bracis-33583 121 16 for for ADP bracis-33583 121 17 each each DET bracis-33583 121 18 neuron neuron NOUN bracis-33583 121 19 . . PUNCT bracis-33583 122 1 the the DET bracis-33583 122 2 prototype prototype NOUN bracis-33583 122 3 of of ADP bracis-33583 122 4 a a DET bracis-33583 122 5 neuron neuron NOUN bracis-33583 122 6 n n PART bracis-33583 122 7 is be AUX bracis-33583 122 8 the the DET bracis-33583 122 9 column column NOUN bracis-33583 122 10 - - PUNCT bracis-33583 122 11 wise wise ADJ bracis-33583 122 12 average average NOUN bracis-33583 122 13 of of ADP bracis-33583 122 14 the the DET bracis-33583 122 15 matrix matrix NOUN bracis-33583 122 16 formed form VERB bracis-33583 122 17 by by ADP bracis-33583 122 18 all all DET bracis-33583 122 19 binary binary ADJ bracis-33583 122 20 label label NOUN bracis-33583 122 21 vectors vector NOUN bracis-33583 122 22 associated associate VERB bracis-33583 122 23 with with ADP bracis-33583 122 24 the the DET bracis-33583 122 25 training training NOUN bracis-33583 122 26 instances instance NOUN bracis-33583 122 27 that that PRON bracis-33583 122 28 had have VERB bracis-33583 122 29 the the DET bracis-33583 122 30 neuron neuron NOUN bracis-33583 122 31 n n CCONJ bracis-33583 122 32 as as ADP bracis-33583 122 33 their their PRON bracis-33583 122 34 winning win VERB bracis-33583 122 35 neuron neuron NOUN bracis-33583 122 36 . . PUNCT bracis-33583 123 1 this this DET bracis-33583 123 2 calculation calculation NOUN bracis-33583 123 3 is be AUX bracis-33583 123 4 presented present VERB bracis-33583 123 5 in in ADP bracis-33583 123 6 eq eq ADP bracis-33583 123 7 . . PUNCT bracis-33583 123 8     SPACE bracis-33583 124 1 4 4 NUM bracis-33583 124 2 , , PUNCT bracis-33583 124 3 where where SCONJ bracis-33583 124 4 \({s_{n \({s_{n NOUN bracis-33583 124 5 , , PUNCT bracis-33583 124 6 j}}\ j}}\ PROPN bracis-33583 124 7 ) ) PUNCT bracis-33583 124 8 is be AUX bracis-33583 124 9 the the DET bracis-33583 124 10 set set NOUN bracis-33583 124 11 of of ADP bracis-33583 124 12 training training NOUN bracis-33583 124 13 instances instance NOUN bracis-33583 124 14 mapped map VERB bracis-33583 124 15 to to ADP bracis-33583 124 16 the the DET bracis-33583 124 17 neuron neuron NOUN bracis-33583 124 18 n n CCONJ bracis-33583 124 19 classified classify VERB bracis-33583 124 20 in in ADP bracis-33583 124 21 the the DET bracis-33583 124 22 class class NOUN bracis-33583 124 23 \(c_j\ \(c_j\ NOUN bracis-33583 124 24 ) ) PUNCT bracis-33583 124 25 and and CCONJ bracis-33583 124 26 \({s_n}\ \({s_n}\ NUM bracis-33583 124 27 ) ) PUNCT bracis-33583 124 28 is be AUX bracis-33583 124 29 the the DET bracis-33583 124 30 total total ADJ bracis-33583 124 31 instances instance NOUN bracis-33583 124 32 mapped map VERB bracis-33583 124 33 to to PART bracis-33583 124 34 neuron neuron VERB bracis-33583 124 35 n. n. NOUN bracis-33583 124 36 $ $ SYM bracis-33583 124 37 $ $ SYM bracis-33583 124 38 \begin{aligned \begin{aligne VERB bracis-33583 124 39 } } PUNCT bracis-33583 124 40 \overline{\textbf{v}}_{i \overline{\textbf{v}}_{i PROPN bracis-33583 124 41 , , PUNCT bracis-33583 124 42 j j NOUN bracis-33583 124 43 } } PUNCT bracis-33583 124 44 = = SYM bracis-33583 124 45 \frac{|s_{n \frac{|s_{n NOUN bracis-33583 124 46 , , PUNCT bracis-33583 124 47 j}|}{|s_n| j}|}{|s_n| PROPN bracis-33583 124 48 } } PUNCT bracis-33583 124 49 \end{aligned}$$ \end{aligned}$$ X bracis-33583 124 50 ( ( PUNCT bracis-33583 124 51 4 4 NUM bracis-33583 124 52 ) ) PUNCT bracis-33583 124 53 the the DET bracis-33583 124 54 prototype prototype NOUN bracis-33583 124 55 vector vector NOUN bracis-33583 124 56 represents represent VERB bracis-33583 124 57 the the DET bracis-33583 124 58 probability probability NOUN bracis-33583 124 59 that that SCONJ bracis-33583 124 60 the the DET bracis-33583 124 61 mapped map VERB bracis-33583 124 62 instance instance NOUN bracis-33583 124 63 \(x_i\ \(x_i\ PROPN bracis-33583 124 64 ) ) PUNCT bracis-33583 124 65 will will AUX bracis-33583 124 66 be be AUX bracis-33583 124 67 classified classify VERB bracis-33583 124 68 in in ADP bracis-33583 124 69 each each PRON bracis-33583 124 70 of of ADP bracis-33583 124 71 the the DET bracis-33583 124 72 problem problem NOUN bracis-33583 124 73 ’s ’s PART bracis-33583 124 74 l l NOUN bracis-33583 124 75 labels label NOUN bracis-33583 124 76 . . PUNCT bracis-33583 125 1 a a DET bracis-33583 125 2 threshold threshold NOUN bracis-33583 125 3 is be AUX bracis-33583 125 4 applied apply VERB bracis-33583 125 5 to to ADP bracis-33583 125 6 the the DET bracis-33583 125 7 prototype prototype NOUN bracis-33583 125 8 vector vector NOUN bracis-33583 125 9 to to PART bracis-33583 125 10 obtain obtain VERB bracis-33583 125 11 a a DET bracis-33583 125 12 final final ADJ bracis-33583 125 13 binary binary ADJ bracis-33583 125 14 prediction prediction NOUN bracis-33583 125 15 . . PUNCT bracis-33583 126 1 the the DET bracis-33583 126 2 common common ADJ bracis-33583 126 3 threshold threshold NOUN bracis-33583 126 4 used use VERB bracis-33583 126 5 is be AUX bracis-33583 126 6 0.5 0.5 NUM bracis-33583 126 7 , , PUNCT bracis-33583 126 8 so so SCONJ bracis-33583 126 9 that that SCONJ bracis-33583 126 10 if if SCONJ bracis-33583 126 11 \(\overline{\textbf{v}}_{i \(\overline{\textbf{v}}_{i NOUN bracis-33583 126 12 , , PUNCT bracis-33583 126 13 j}\ j}\ PROPN bracis-33583 126 14 ) ) PUNCT bracis-33583 126 15 is be AUX bracis-33583 126 16 greater great ADJ bracis-33583 126 17 or or CCONJ bracis-33583 126 18 equal equal ADJ bracis-33583 126 19 to to ADP bracis-33583 126 20 0.5 0.5 NUM bracis-33583 126 21 , , PUNCT bracis-33583 126 22 \(x_i\ \(x_i\ NOUN bracis-33583 126 23 ) ) PUNCT bracis-33583 126 24 is be AUX bracis-33583 126 25 classified classify VERB bracis-33583 126 26 in in ADP bracis-33583 126 27 class class NOUN bracis-33583 126 28 \(c_j\ \(c_j\ NOUN bracis-33583 126 29 ) ) PUNCT bracis-33583 126 30 . . PUNCT bracis-33583 127 1 figure figure NOUN bracis-33583 127 2     SPACE bracis-33583 127 3 2 2 NUM bracis-33583 127 4 shows show VERB bracis-33583 127 5 an an DET bracis-33583 127 6 example example NOUN bracis-33583 127 7 of of ADP bracis-33583 127 8 applying apply VERB bracis-33583 127 9 a a DET bracis-33583 127 10 threshold threshold NOUN bracis-33583 127 11 to to ADP bracis-33583 127 12 a a DET bracis-33583 127 13 prototype prototype NOUN bracis-33583 127 14 vector vector NOUN bracis-33583 127 15 . . PUNCT bracis-33583 128 1 fig fig NOUN bracis-33583 128 2 . . PUNCT bracis-33583 129 1 2 2 X bracis-33583 129 2 . . X bracis-33583 129 3 obtaining obtain VERB bracis-33583 129 4 a a DET bracis-33583 129 5 final final ADJ bracis-33583 129 6 prediction prediction NOUN bracis-33583 129 7 by by ADP bracis-33583 129 8 applying apply VERB bracis-33583 129 9 a a DET bracis-33583 129 10 threshold threshold NOUN bracis-33583 129 11 to to ADP bracis-33583 129 12 a a DET bracis-33583 129 13 prototype prototype NOUN bracis-33583 129 14 vector vector NOUN bracis-33583 129 15 full full ADJ bracis-33583 129 16 size size NOUN bracis-33583 129 17 image image NOUN bracis-33583 129 18 4 4 NUM bracis-33583 129 19 growing grow VERB bracis-33583 129 20 self self NOUN bracis-33583 129 21 - - PUNCT bracis-33583 129 22 organizing organize VERB bracis-33583 129 23 maps map NOUN bracis-33583 129 24 for for ADP bracis-33583 129 25     SPACE bracis-33583 129 26 multi multi ADJ bracis-33583 129 27 - - ADJ bracis-33583 129 28 label label ADJ bracis-33583 129 29 classification classification NOUN bracis-33583 129 30 this this DET bracis-33583 129 31 section section NOUN bracis-33583 129 32 introduces introduce VERB bracis-33583 129 33 the the DET bracis-33583 129 34 growing grow VERB bracis-33583 129 35 self self NOUN bracis-33583 129 36 - - PUNCT bracis-33583 129 37 organizing organize VERB bracis-33583 129 38 maps map NOUN bracis-33583 129 39 ( ( PUNCT bracis-33583 129 40 gsom gsom PROPN bracis-33583 129 41 ) ) PUNCT bracis-33583 129 42 and and CCONJ bracis-33583 129 43 our our PRON bracis-33583 129 44 proposal proposal NOUN bracis-33583 129 45 for for ADP bracis-33583 129 46 a a DET bracis-33583 129 47 multi multi ADJ bracis-33583 129 48 - - ADJ bracis-33583 129 49 label label ADJ bracis-33583 129 50 classification classification NOUN bracis-33583 129 51 method method NOUN bracis-33583 129 52 called call VERB bracis-33583 129 53 growing grow VERB bracis-33583 129 54 self self NOUN bracis-33583 129 55 - - PUNCT bracis-33583 129 56 organizing organize VERB bracis-33583 129 57 maps map NOUN bracis-33583 129 58 multi multi ADJ bracis-33583 129 59 - - ADJ bracis-33583 129 60 label label ADJ bracis-33583 129 61 learning learning NOUN bracis-33583 129 62 ( ( PUNCT bracis-33583 129 63 gsom gsom PROPN bracis-33583 129 64 - - PUNCT bracis-33583 129 65 mll mll PROPN bracis-33583 129 66 ) ) PUNCT bracis-33583 129 67 . . PUNCT bracis-33583 130 1 4.1 4.1 NUM bracis-33583 130 2 growing grow VERB bracis-33583 130 3 self self NOUN bracis-33583 130 4 - - PUNCT bracis-33583 130 5 organizing organize VERB bracis-33583 130 6 maps map NOUN bracis-33583 130 7 the the DET bracis-33583 130 8 growing grow VERB bracis-33583 130 9 self self NOUN bracis-33583 130 10 - - PUNCT bracis-33583 130 11 organizing organize VERB bracis-33583 130 12 maps map NOUN bracis-33583 130 13     SPACE bracis-33583 131 1 [ [ X bracis-33583 131 2 6 6 NUM bracis-33583 131 3 ] ] PUNCT bracis-33583 131 4 is be AUX bracis-33583 131 5 an an DET bracis-33583 131 6 extension extension NOUN bracis-33583 131 7 of of ADP bracis-33583 131 8 the the DET bracis-33583 131 9 kohonen kohonen PROPN bracis-33583 131 10 maps map NOUN bracis-33583 131 11 able able ADJ bracis-33583 131 12 to to PART bracis-33583 131 13 increase increase VERB bracis-33583 131 14 its its PRON bracis-33583 131 15 number number NOUN bracis-33583 131 16 of of ADP bracis-33583 131 17 neurons neuron NOUN bracis-33583 131 18 during during ADP bracis-33583 131 19 training training NOUN bracis-33583 131 20 based base VERB bracis-33583 131 21 on on ADP bracis-33583 131 22 the the DET bracis-33583 131 23 characteristics characteristic NOUN bracis-33583 131 24 of of ADP bracis-33583 131 25 the the DET bracis-33583 131 26 input input NOUN bracis-33583 131 27 data datum NOUN bracis-33583 131 28 . . PUNCT bracis-33583 132 1 the the DET bracis-33583 132 2 algorithm algorithm NOUN bracis-33583 132 3 has have VERB bracis-33583 132 4 two two NUM bracis-33583 132 5 phases phase NOUN bracis-33583 132 6 : : PUNCT bracis-33583 132 7 the the DET bracis-33583 132 8 growing grow VERB bracis-33583 132 9 phase phase NOUN bracis-33583 132 10 , , PUNCT bracis-33583 132 11 where where SCONJ bracis-33583 132 12 new new ADJ bracis-33583 132 13 neurons neuron NOUN bracis-33583 132 14 are be AUX bracis-33583 132 15 added add VERB bracis-33583 132 16 to to ADP bracis-33583 132 17 the the DET bracis-33583 132 18 map map NOUN bracis-33583 132 19 , , PUNCT bracis-33583 132 20 and and CCONJ bracis-33583 132 21 the the DET bracis-33583 132 22 smooth smooth ADJ bracis-33583 132 23 phase phase NOUN bracis-33583 132 24 , , PUNCT bracis-33583 132 25 where where SCONJ bracis-33583 132 26 the the DET bracis-33583 132 27 weights weight NOUN bracis-33583 132 28 of of ADP bracis-33583 132 29 neurons neuron NOUN bracis-33583 132 30 are be AUX bracis-33583 132 31 faceted facete VERB bracis-33583 132 32 by by ADP bracis-33583 132 33 their their PRON bracis-33583 132 34 neighbors neighbor NOUN bracis-33583 132 35 . . PUNCT bracis-33583 133 1 the the DET bracis-33583 133 2 algorithm algorithm NOUN bracis-33583 133 3 has have VERB bracis-33583 133 4 a a DET bracis-33583 133 5 hyperparameter hyperparameter NOUN bracis-33583 133 6 called call VERB bracis-33583 133 7 spread spread ADJ bracis-33583 133 8 factor factor NOUN bracis-33583 133 9 sf sf PROPN bracis-33583 133 10 , , PUNCT bracis-33583 133 11 \(0 \(0 PROPN bracis-33583 133 12 \le \le PROPN bracis-33583 133 13 sf sf PROPN bracis-33583 133 14 \le \le PROPN bracis-33583 133 15 1\ 1\ NUM bracis-33583 133 16 ) ) PUNCT bracis-33583 133 17 , , PUNCT bracis-33583 133 18 which which PRON bracis-33583 133 19 controls control VERB bracis-33583 133 20 the the DET bracis-33583 133 21 number number NOUN bracis-33583 133 22 of of ADP bracis-33583 133 23 grid grid NOUN bracis-33583 133 24 nodes node NOUN bracis-33583 133 25 . . PUNCT bracis-33583 134 1 lower low ADJ bracis-33583 134 2 sf sf PROPN bracis-33583 134 3 values value NOUN bracis-33583 134 4 create create VERB bracis-33583 134 5 more more ADV bracis-33583 134 6 populated populated ADJ bracis-33583 134 7 maps map NOUN bracis-33583 134 8 , , PUNCT bracis-33583 134 9 while while SCONJ bracis-33583 134 10 higher high ADJ bracis-33583 134 11 values value NOUN bracis-33583 134 12 create create VERB bracis-33583 134 13 more more ADV bracis-33583 134 14 concentrated concentrated ADJ bracis-33583 134 15 maps map NOUN bracis-33583 134 16 . . PUNCT bracis-33583 135 1 4.2 4.2 NUM bracis-33583 135 2 growing grow VERB bracis-33583 135 3 phase phase NOUN bracis-33583 135 4 the the DET bracis-33583 135 5 growing grow VERB bracis-33583 135 6 phase phase NOUN bracis-33583 135 7 is be AUX bracis-33583 135 8 responsible responsible ADJ bracis-33583 135 9 for for ADP bracis-33583 135 10 creating create VERB bracis-33583 135 11 new new ADJ bracis-33583 135 12 nodes node NOUN bracis-33583 135 13 on on ADP bracis-33583 135 14 the the DET bracis-33583 135 15 map map NOUN bracis-33583 135 16 by by ADP bracis-33583 135 17 means mean NOUN bracis-33583 135 18 of of ADP bracis-33583 135 19 a a DET bracis-33583 135 20 growth growth NOUN bracis-33583 135 21 threshold threshold NOUN bracis-33583 135 22 ( ( PUNCT bracis-33583 135 23 gt gt PROPN bracis-33583 135 24 ) ) PUNCT bracis-33583 135 25 , , PUNCT bracis-33583 135 26 a a DET bracis-33583 135 27 spread spread ADJ bracis-33583 135 28 factor factor NOUN bracis-33583 135 29 ( ( PUNCT bracis-33583 135 30 sf sf PROPN bracis-33583 135 31 ) ) PUNCT bracis-33583 135 32 , , PUNCT bracis-33583 135 33 and and CCONJ bracis-33583 135 34 the the DET bracis-33583 135 35 dimension dimension NOUN bracis-33583 135 36 d d PROPN bracis-33583 135 37 of of ADP bracis-33583 135 38 the the DET bracis-33583 135 39 training training NOUN bracis-33583 135 40 data datum NOUN bracis-33583 135 41 . . PUNCT bracis-33583 136 1 this this PRON bracis-33583 136 2 can can AUX bracis-33583 136 3 be be AUX bracis-33583 136 4 seen see VERB bracis-33583 136 5 in in ADP bracis-33583 136 6 eq eq NOUN bracis-33583 136 7 . . PUNCT bracis-33583 136 8     SPACE bracis-33583 137 1 5 5 NUM bracis-33583 137 2 . . PUNCT bracis-33583 137 3 $ $ SYM bracis-33583 137 4 $ $ SYM bracis-33583 137 5 \begin{aligned \begin{aligne VERB bracis-33583 137 6 } } PUNCT bracis-33583 137 7 gt gt PUNCT bracis-33583 138 1 = = PUNCT bracis-33583 138 2 d d NOUN bracis-33583 138 3 \times \time NOUN bracis-33583 138 4 \log \log PROPN bracis-33583 138 5 { { PUNCT bracis-33583 138 6 sf sf PROPN bracis-33583 138 7 } } PUNCT bracis-33583 138 8 \end{aligned}$$ \end{aligned}$$ X bracis-33583 138 9 ( ( PUNCT bracis-33583 138 10 5 5 NUM bracis-33583 138 11 ) ) PUNCT bracis-33583 138 12 the the DET bracis-33583 138 13 euclidean euclidean ADJ bracis-33583 138 14 distance distance NOUN bracis-33583 138 15 ( ( PUNCT bracis-33583 138 16 eq eq NOUN bracis-33583 138 17 . . PUNCT bracis-33583 138 18     SPACE bracis-33583 138 19 1 1 X bracis-33583 138 20 ) ) PUNCT bracis-33583 138 21 is be AUX bracis-33583 138 22 used use VERB bracis-33583 138 23 for for ADP bracis-33583 138 24 each each DET bracis-33583 138 25 input input NOUN bracis-33583 138 26 training training NOUN bracis-33583 138 27 instance instance NOUN bracis-33583 138 28 to to PART bracis-33583 138 29 find find VERB bracis-33583 138 30 the the DET bracis-33583 138 31 corresponding corresponding ADJ bracis-33583 138 32 closest close ADJ bracis-33583 138 33 neuron neuron NOUN bracis-33583 138 34 . . PUNCT bracis-33583 139 1 after after ADP bracis-33583 139 2 this this DET bracis-33583 139 3 winning winning NOUN bracis-33583 139 4 neuron neuron NOUN bracis-33583 139 5 i i PRON bracis-33583 139 6 is be AUX bracis-33583 139 7 found find VERB bracis-33583 139 8 , , PUNCT bracis-33583 139 9 the the DET bracis-33583 139 10 weights weight NOUN bracis-33583 139 11 are be AUX bracis-33583 139 12 adjusted adjust VERB bracis-33583 139 13 for for ADP bracis-33583 139 14 all all DET bracis-33583 139 15 neurons neuron NOUN bracis-33583 139 16 j j PROPN bracis-33583 139 17 of of ADP bracis-33583 139 18 the the DET bracis-33583 139 19 grid grid NOUN bracis-33583 139 20 in in ADP bracis-33583 139 21 the the DET bracis-33583 139 22 neighborhood neighborhood NOUN bracis-33583 139 23 influence influence NOUN bracis-33583 139 24 ( ( PUNCT bracis-33583 139 25 \(h_{j \(h_{j NOUN bracis-33583 139 26 , , PUNCT bracis-33583 139 27 i}\ i}\ NUM bracis-33583 139 28 ) ) PUNCT bracis-33583 139 29 ) ) PUNCT bracis-33583 139 30 of of ADP bracis-33583 139 31 neuron neuron PROPN bracis-33583 139 32 i. i. PROPN bracis-33583 139 33 this this DET bracis-33583 139 34 adjustment adjustment NOUN bracis-33583 139 35 procedure procedure NOUN bracis-33583 139 36 is be AUX bracis-33583 139 37 the the DET bracis-33583 139 38 same same ADJ bracis-33583 139 39 one one NOUN bracis-33583 139 40 used use VERB bracis-33583 139 41 in in ADP bracis-33583 139 42 the the DET bracis-33583 139 43 original original ADJ bracis-33583 139 44 som som NOUN bracis-33583 139 45 ( ( PUNCT bracis-33583 139 46 eqs eqs X bracis-33583 139 47 . . PUNCT bracis-33583 139 48     SPACE bracis-33583 139 49 1 1 NUM bracis-33583 139 50 , , PUNCT bracis-33583 139 51 2 2 NUM bracis-33583 139 52 and and CCONJ bracis-33583 139 53 3 3 NUM bracis-33583 139 54 ) ) PUNCT bracis-33583 139 55 . . PUNCT bracis-33583 140 1 as as ADP bracis-33583 140 2 an an DET bracis-33583 140 3 extension extension NOUN bracis-33583 140 4 of of ADP bracis-33583 140 5 soms som NOUN bracis-33583 140 6 , , PUNCT bracis-33583 140 7 each each DET bracis-33583 140 8 neuron neuron NOUN bracis-33583 140 9 in in ADP bracis-33583 140 10 gsom gsom PROPN bracis-33583 140 11 has have VERB bracis-33583 140 12 an an DET bracis-33583 140 13 associated associated ADJ bracis-33583 140 14 error error NOUN bracis-33583 140 15 value value NOUN bracis-33583 140 16 e e NOUN bracis-33583 140 17 used use VERB bracis-33583 140 18 to to PART bracis-33583 140 19 compute compute VERB bracis-33583 140 20 the the DET bracis-33583 140 21 accumulated accumulate VERB bracis-33583 140 22 error error NOUN bracis-33583 140 23 in in ADP bracis-33583 140 24 the the DET bracis-33583 140 25 neuron neuron NOUN bracis-33583 140 26 . . PUNCT bracis-33583 141 1 this this PRON bracis-33583 141 2 is be AUX bracis-33583 141 3 shown show VERB bracis-33583 141 4 in in ADP bracis-33583 141 5 eq eq PROPN bracis-33583 141 6 . . PUNCT bracis-33583 141 7     SPACE bracis-33583 141 8 6 6 NUM bracis-33583 141 9 , , PUNCT bracis-33583 141 10 where where SCONJ bracis-33583 141 11 \(\textbf{x}\ \(\textbf{x}\ X bracis-33583 141 12 ) ) PUNCT bracis-33583 141 13 is be AUX bracis-33583 141 14 the the DET bracis-33583 141 15 input input NOUN bracis-33583 141 16 sample sample NOUN bracis-33583 141 17 , , PUNCT bracis-33583 141 18 and and CCONJ bracis-33583 141 19 a a PRON bracis-33583 141 20 is be AUX bracis-33583 141 21 the the DET bracis-33583 141 22 number number NOUN bracis-33583 141 23 of of ADP bracis-33583 141 24 attributes attribute NOUN bracis-33583 141 25 . . PUNCT bracis-33583 142 1 $ $ SYM bracis-33583 142 2 $ $ SYM bracis-33583 142 3 \begin{aligned \begin{aligne VERB bracis-33583 142 4 } } PUNCT bracis-33583 142 5 e_j(t+1 e_j(t+1 NOUN bracis-33583 142 6 ) ) PUNCT bracis-33583 143 1 = = SYM bracis-33583 143 2 e_j(t e_j(t X bracis-33583 143 3 ) ) PUNCT bracis-33583 144 1 + + CCONJ bracis-33583 144 2 \sqrt{\sum \sqrt{\sum X bracis-33583 144 3 _ _ PRON bracis-33583 144 4 { { PUNCT bracis-33583 144 5 i=1}^{a i=1}^{a NOUN bracis-33583 144 6 } } PUNCT bracis-33583 144 7 { { PUNCT bracis-33583 144 8 ( ( PUNCT bracis-33583 144 9 x_i x_i PUNCT bracis-33583 144 10 w_{j w_{j NOUN bracis-33583 144 11 , , PUNCT bracis-33583 144 12 i})^2 i})^2 PROPN bracis-33583 144 13 } } PUNCT bracis-33583 144 14 } } PUNCT bracis-33583 144 15 \end{aligned}$$ \end{aligned}$$ X bracis-33583 144 16 ( ( PUNCT bracis-33583 144 17 6 6 NUM bracis-33583 144 18 ) ) PUNCT bracis-33583 144 19 the the DET bracis-33583 144 20 largest large ADJ bracis-33583 144 21 error error NOUN bracis-33583 144 22 value value NOUN bracis-33583 144 23 e e NOUN bracis-33583 144 24 is be AUX bracis-33583 144 25 called call VERB bracis-33583 144 26 \(h_{e_{rr}}\ \(h_{e_{rr}}\ ADV bracis-33583 144 27 ) ) PUNCT bracis-33583 144 28 , , PUNCT bracis-33583 144 29 and and CCONJ bracis-33583 144 30 is be AUX bracis-33583 144 31 updated update VERB bracis-33583 144 32 every every DET bracis-33583 144 33 time time NOUN bracis-33583 144 34 a a DET bracis-33583 144 35 new new ADJ bracis-33583 144 36 input input NOUN bracis-33583 144 37 instance instance NOUN bracis-33583 144 38 is be AUX bracis-33583 144 39 given give VERB bracis-33583 144 40 to to ADP bracis-33583 144 41 the the DET bracis-33583 144 42 model model NOUN bracis-33583 144 43 . . PUNCT bracis-33583 145 1 if if SCONJ bracis-33583 145 2 \(h_{e_{rr \(h_{e_{rr PROPN bracis-33583 145 3 } } PUNCT bracis-33583 145 4 } } PUNCT bracis-33583 145 5 > > X bracis-33583 145 6 gt\ gt\ PROPN bracis-33583 145 7 ) ) PUNCT bracis-33583 145 8 , , PUNCT bracis-33583 145 9 the the DET bracis-33583 145 10 model model NOUN bracis-33583 145 11 creates create VERB bracis-33583 145 12 a a DET bracis-33583 145 13 new new ADJ bracis-33583 145 14 node node NOUN bracis-33583 145 15 and and CCONJ bracis-33583 145 16 sets set VERB bracis-33583 145 17 \(h_{e_{rr \(h_{e_{rr NOUN bracis-33583 145 18 } } PUNCT bracis-33583 145 19 } } PUNCT bracis-33583 145 20 = = SYM bracis-33583 145 21 gt/2\ gt/2\ NOUN bracis-33583 145 22 ) ) PUNCT bracis-33583 145 23 . . PUNCT bracis-33583 146 1 4.3 4.3 NUM bracis-33583 146 2 node node NOUN bracis-33583 146 3 generation generation NOUN bracis-33583 146 4 a a DET bracis-33583 146 5 new new ADJ bracis-33583 146 6 node node NOUN bracis-33583 146 7 will will AUX bracis-33583 146 8 always always ADV bracis-33583 146 9 be be AUX bracis-33583 146 10 connected connect VERB bracis-33583 146 11 to to ADP bracis-33583 146 12 a a DET bracis-33583 146 13 boundary boundary ADJ bracis-33583 146 14 node node NOUN bracis-33583 146 15 , , PUNCT bracis-33583 146 16 a a DET bracis-33583 146 17 neuron neuron NOUN bracis-33583 146 18 with with ADP bracis-33583 146 19 at at ADV bracis-33583 146 20 least least ADJ bracis-33583 146 21 one one NUM bracis-33583 146 22 of of ADP bracis-33583 146 23 its its PRON bracis-33583 146 24 four four NUM bracis-33583 146 25 immediate immediate ADJ bracis-33583 146 26 neighboring neighboring NOUN bracis-33583 146 27 positions position NOUN bracis-33583 146 28 free free ADJ bracis-33583 146 29 . . PUNCT bracis-33583 147 1 figure figure NOUN bracis-33583 147 2     SPACE bracis-33583 147 3 3 3 NUM bracis-33583 147 4 shows show VERB bracis-33583 147 5 the the DET bracis-33583 147 6 process process NOUN bracis-33583 147 7 of of ADP bracis-33583 147 8 a a DET bracis-33583 147 9 node node ADJ bracis-33583 147 10 generation generation NOUN bracis-33583 147 11 , , PUNCT bracis-33583 147 12 being be AUX bracis-33583 147 13 ( ( PUNCT bracis-33583 147 14 i i NOUN bracis-33583 147 15 ) ) PUNCT bracis-33583 147 16 the the DET bracis-33583 147 17 original original ADJ bracis-33583 147 18 grid grid NOUN bracis-33583 147 19 , , PUNCT bracis-33583 147 20 ( ( PUNCT bracis-33583 147 21 ii ii NOUN bracis-33583 147 22 ) ) PUNCT bracis-33583 147 23 the the DET bracis-33583 147 24 grid grid NOUN bracis-33583 147 25 where where SCONJ bracis-33583 147 26 a a DET bracis-33583 147 27 neuron neuron NOUN bracis-33583 147 28 has have VERB bracis-33583 147 29 an an DET bracis-33583 147 30 error error NOUN bracis-33583 147 31 larger large ADJ bracis-33583 147 32 than than ADP bracis-33583 147 33 gt gt PROPN bracis-33583 147 34 , , PUNCT bracis-33583 147 35 and and CCONJ bracis-33583 147 36 ( ( PUNCT bracis-33583 147 37 iii iii X bracis-33583 147 38 ) ) PUNCT bracis-33583 147 39 the the DET bracis-33583 147 40 two two NUM bracis-33583 147 41 possible possible ADJ bracis-33583 147 42 positions position NOUN bracis-33583 147 43 where where SCONJ bracis-33583 147 44 a a DET bracis-33583 147 45 new new ADJ bracis-33583 147 46 neuron neuron NOUN bracis-33583 147 47 can can AUX bracis-33583 147 48 be be AUX bracis-33583 147 49 inserted insert VERB bracis-33583 147 50 in in ADP bracis-33583 147 51 the the DET bracis-33583 147 52 grid grid NOUN bracis-33583 147 53 . . PUNCT bracis-33583 148 1 fig fig NOUN bracis-33583 148 2 . . PUNCT bracis-33583 149 1 3.adapted 3.adapted NUM bracis-33583 149 2 from from ADP bracis-33583 149 3 alahakoon alahakoon PROPN bracis-33583 149 4 et et PROPN bracis-33583 149 5 al al PROPN bracis-33583 149 6 . . PUNCT bracis-33583 149 7     SPACE bracis-33583 150 1 [ [ X bracis-33583 150 2 6 6 NUM bracis-33583 150 3 ] ] PUNCT bracis-33583 150 4 insertion insertion NOUN bracis-33583 150 5 of of ADP bracis-33583 150 6 a a DET bracis-33583 150 7 new new ADJ bracis-33583 150 8 neuron neuron NOUN bracis-33583 150 9 in in ADP bracis-33583 150 10 a a DET bracis-33583 150 11 grid grid NOUN bracis-33583 150 12 . . PUNCT bracis-33583 151 1 full full ADJ bracis-33583 151 2 size size NOUN bracis-33583 151 3 image image NOUN bracis-33583 151 4 the the DET bracis-33583 151 5 weights weight NOUN bracis-33583 151 6 vector vector NOUN bracis-33583 151 7 \(\textbf{w}_{new}\ \(\textbf{w}_{new}\ PROPN bracis-33583 151 8 ) ) PUNCT bracis-33583 151 9 of of ADP bracis-33583 151 10 a a DET bracis-33583 151 11 new new ADJ bracis-33583 151 12 neuron neuron NOUN bracis-33583 151 13 is be AUX bracis-33583 151 14 based base VERB bracis-33583 151 15 on on ADP bracis-33583 151 16 the the DET bracis-33583 151 17 weight weight NOUN bracis-33583 151 18 vectors vector NOUN bracis-33583 151 19 \(\textbf{w}1\ \(\textbf{w}1\ VERB bracis-33583 151 20 ) ) PUNCT bracis-33583 151 21 and and CCONJ bracis-33583 151 22 \(\textbf{w}2\ \(\textbf{w}2\ NUM bracis-33583 151 23 ) ) PUNCT bracis-33583 151 24 of of ADP bracis-33583 151 25 its its PRON bracis-33583 151 26 neighbors neighbor NOUN bracis-33583 151 27 so so SCONJ bracis-33583 151 28 that that SCONJ bracis-33583 151 29 certain certain ADJ bracis-33583 151 30 regions region NOUN bracis-33583 151 31 have have VERB bracis-33583 151 32 similar similar ADJ bracis-33583 151 33 characteristics characteristic NOUN bracis-33583 151 34 . . PUNCT bracis-33583 152 1 figure figure NOUN bracis-33583 152 2     SPACE bracis-33583 152 3 4 4 NUM bracis-33583 152 4 illustrates illustrate VERB bracis-33583 152 5 the the DET bracis-33583 152 6 four four NUM bracis-33583 152 7 cases case NOUN bracis-33583 152 8 that that PRON bracis-33583 152 9 dictate dictate VERB bracis-33583 152 10 the the DET bracis-33583 152 11 weight weight NOUN bracis-33583 152 12 initialization initialization NOUN bracis-33583 152 13 of of ADP bracis-33583 152 14 a a DET bracis-33583 152 15 new new ADJ bracis-33583 152 16 neuron neuron NOUN bracis-33583 152 17 . . PUNCT bracis-33583 153 1 all all DET bracis-33583 153 2 comparisons comparison NOUN bracis-33583 153 3 and and CCONJ bracis-33583 153 4 operations operation NOUN bracis-33583 153 5 in in ADP bracis-33583 153 6 the the DET bracis-33583 153 7 four four NUM bracis-33583 153 8 cases case NOUN bracis-33583 153 9 are be AUX bracis-33583 153 10 executed execute VERB bracis-33583 153 11 individually individually ADV bracis-33583 153 12 for for ADP bracis-33583 153 13 each each DET bracis-33583 153 14 weight weight NOUN bracis-33583 153 15 of of ADP bracis-33583 153 16 the the DET bracis-33583 153 17 vectors vector NOUN bracis-33583 153 18 \(\textbf{w}1\ \(\textbf{w}1\ VERB bracis-33583 153 19 ) ) PUNCT bracis-33583 153 20 and and CCONJ bracis-33583 153 21 \(\textbf{w}2\ \(\textbf{w}2\ ADJ bracis-33583 153 22 ) ) PUNCT bracis-33583 153 23 . . PUNCT bracis-33583 154 1 1 1 X bracis-33583 154 2 . . X bracis-33583 154 3 the the DET bracis-33583 154 4 new new ADJ bracis-33583 154 5 neuron neuron NOUN bracis-33583 154 6 has have VERB bracis-33583 154 7 two two NUM bracis-33583 154 8 consecutive consecutive ADJ bracis-33583 154 9 older old ADJ bracis-33583 154 10 nodes node NOUN bracis-33583 154 11 on on ADP bracis-33583 154 12 one one NUM bracis-33583 154 13 of of ADP bracis-33583 154 14 its its PRON bracis-33583 154 15 sides side NOUN bracis-33583 154 16 ( ( PUNCT bracis-33583 154 17 fig fig NOUN bracis-33583 154 18 . . PUNCT bracis-33583 154 19     SPACE bracis-33583 154 20 4(a 4(a NUM bracis-33583 154 21 ) ) PUNCT bracis-33583 154 22 ) ) PUNCT bracis-33583 154 23 . . PUNCT bracis-33583 155 1 in in ADP bracis-33583 155 2 this this DET bracis-33583 155 3 case case NOUN bracis-33583 155 4 , , PUNCT bracis-33583 155 5 the the DET bracis-33583 155 6 following follow VERB bracis-33583 155 7 rules rule NOUN bracis-33583 155 8 are be AUX bracis-33583 155 9 applied apply VERB bracis-33583 155 10 : : PUNCT bracis-33583 155 11 if if SCONJ bracis-33583 155 12 \(w2 \(w2 X bracis-33583 155 13 > > X bracis-33583 155 14 w1\ w1\ PROPN bracis-33583 155 15 ) ) PUNCT bracis-33583 155 16 , , PUNCT bracis-33583 155 17 then then ADV bracis-33583 155 18 \(w_{new \(w_{new ADJ bracis-33583 155 19 } } PUNCT bracis-33583 155 20 = = SYM bracis-33583 155 21 w1 w1 NOUN bracis-33583 155 22 ( ( PUNCT bracis-33583 155 23 w2 w2 NOUN bracis-33583 155 24 w1)\ w1)\ PROPN bracis-33583 155 25 ) ) PUNCT bracis-33583 155 26 ; ; PUNCT bracis-33583 155 27 if if SCONJ bracis-33583 155 28 \(w1 \(w1 ADP bracis-33583 155 29 > > X bracis-33583 155 30 w2\ w2\ PROPN bracis-33583 155 31 ) ) PUNCT bracis-33583 155 32 , , PUNCT bracis-33583 155 33 then then ADV bracis-33583 155 34 \(w_{new \(w_{new ADJ bracis-33583 155 35 } } PUNCT bracis-33583 155 36 = = SYM bracis-33583 155 37 w1 w1 NOUN bracis-33583 155 38 + + CCONJ bracis-33583 155 39 ( ( PUNCT bracis-33583 155 40 w1 w1 NOUN bracis-33583 155 41 w2)\ w2)\ NOUN bracis-33583 155 42 ) ) PUNCT bracis-33583 155 43 . . PUNCT bracis-33583 156 1 2 2 X bracis-33583 156 2 . . X bracis-33583 156 3 the the DET bracis-33583 156 4 new new ADJ bracis-33583 156 5 neuron neuron NOUN bracis-33583 156 6 is be AUX bracis-33583 156 7 in in ADP bracis-33583 156 8 between between ADP bracis-33583 156 9 two two NUM bracis-33583 156 10 older old ADJ bracis-33583 156 11 nodes node NOUN bracis-33583 156 12 ( ( PUNCT bracis-33583 156 13 fig fig NOUN bracis-33583 156 14 . . PUNCT bracis-33583 156 15     SPACE bracis-33583 156 16 4(b 4(b NUM bracis-33583 156 17 ) ) PUNCT bracis-33583 156 18 ) ) PUNCT bracis-33583 156 19 . . PUNCT bracis-33583 157 1 in in ADP bracis-33583 157 2 this this DET bracis-33583 157 3 case case NOUN bracis-33583 157 4 : : PUNCT bracis-33583 157 5 \(w_{new \(w_{new ADJ bracis-33583 157 6 } } PUNCT bracis-33583 157 7 = = SYM bracis-33583 157 8 \frac{w1 \frac{w1 NOUN bracis-33583 157 9 + + CCONJ bracis-33583 157 10 w2}{2}\ w2}{2}\ NOUN bracis-33583 157 11 ) ) PUNCT bracis-33583 157 12 . . PUNCT bracis-33583 158 1 3 3 X bracis-33583 158 2 . . X bracis-33583 158 3 the the DET bracis-33583 158 4 new new ADJ bracis-33583 158 5 neuron neuron NOUN bracis-33583 158 6 has have VERB bracis-33583 158 7 only only ADV bracis-33583 158 8 one one NUM bracis-33583 158 9 direct direct ADJ bracis-33583 158 10 neighbor neighbor NOUN bracis-33583 158 11 : : PUNCT bracis-33583 158 12 an an DET bracis-33583 158 13 older old ADJ bracis-33583 158 14 neuron neuron NOUN bracis-33583 158 15 ( ( PUNCT bracis-33583 158 16 fig fig NOUN bracis-33583 158 17 . . PUNCT bracis-33583 158 18     SPACE bracis-33583 158 19 4(c 4(c NUM bracis-33583 158 20 ) ) PUNCT bracis-33583 158 21 ) ) PUNCT bracis-33583 158 22 . . PUNCT bracis-33583 159 1 in in ADP bracis-33583 159 2 this this DET bracis-33583 159 3 case case NOUN bracis-33583 159 4 , , PUNCT bracis-33583 159 5 the the DET bracis-33583 159 6 following follow VERB bracis-33583 159 7 rules rule NOUN bracis-33583 159 8 are be AUX bracis-33583 159 9 applied apply VERB bracis-33583 159 10 : : PUNCT bracis-33583 159 11 if if SCONJ bracis-33583 159 12 \(w2 \(w2 X bracis-33583 159 13 > > X bracis-33583 159 14 w1\ w1\ PROPN bracis-33583 159 15 ) ) PUNCT bracis-33583 159 16 , , PUNCT bracis-33583 159 17 then then ADV bracis-33583 159 18 \(w_{new \(w_{new ADJ bracis-33583 159 19 } } PUNCT bracis-33583 159 20 = = SYM bracis-33583 159 21 w1 w1 NOUN bracis-33583 159 22 ( ( PUNCT bracis-33583 159 23 w2 w2 NOUN bracis-33583 159 24 w1)\ w1)\ PROPN bracis-33583 159 25 ) ) PUNCT bracis-33583 159 26 ; ; PUNCT bracis-33583 159 27 if if SCONJ bracis-33583 159 28 \(w1 \(w1 ADP bracis-33583 159 29 > > X bracis-33583 159 30 w2\ w2\ PROPN bracis-33583 159 31 ) ) PUNCT bracis-33583 159 32 , , PUNCT bracis-33583 159 33 then then ADV bracis-33583 159 34 \(w_{new \(w_{new ADJ bracis-33583 159 35 } } PUNCT bracis-33583 159 36 = = SYM bracis-33583 159 37 w1 w1 NOUN bracis-33583 159 38 + + CCONJ bracis-33583 159 39 ( ( PUNCT bracis-33583 159 40 w1 w1 NOUN bracis-33583 159 41 w2)\ w2)\ NOUN bracis-33583 159 42 ) ) PUNCT bracis-33583 159 43 . . PUNCT bracis-33583 160 1 4 4 X bracis-33583 160 2 . . X bracis-33583 160 3 the the DET bracis-33583 160 4 new new ADJ bracis-33583 160 5 neuron neuron NOUN bracis-33583 160 6 has have VERB bracis-33583 160 7 only only ADV bracis-33583 160 8 one one NUM bracis-33583 160 9 direct direct ADJ bracis-33583 160 10 neighbor neighbor NOUN bracis-33583 160 11 , , PUNCT bracis-33583 160 12 which which PRON bracis-33583 160 13 is be AUX bracis-33583 160 14 an an DET bracis-33583 160 15 old old ADJ bracis-33583 160 16 isolated isolated ADJ bracis-33583 160 17 neuron neuron NOUN bracis-33583 160 18 ( ( PUNCT bracis-33583 160 19 fig fig NOUN bracis-33583 160 20 . . PUNCT bracis-33583 160 21     SPACE bracis-33583 160 22 4(d 4(d NUM bracis-33583 160 23 ) ) PUNCT bracis-33583 160 24 ) ) PUNCT bracis-33583 160 25 . . PUNCT bracis-33583 161 1 in in ADP bracis-33583 161 2 this this DET bracis-33583 161 3 case case NOUN bracis-33583 161 4 : : PUNCT bracis-33583 161 5 \(w_{new \(w_{new ADJ bracis-33583 161 6 } } PUNCT bracis-33583 161 7 = = SYM bracis-33583 161 8 ( ( PUNCT bracis-33583 161 9 r1 r1 PROPN bracis-33583 161 10 + + CCONJ bracis-33583 161 11 r2)/2\ r2)/2\ PROPN bracis-33583 161 12 ) ) PUNCT bracis-33583 161 13 , , PUNCT bracis-33583 161 14 where where SCONJ bracis-33583 161 15 r1 r1 PROPN bracis-33583 161 16 and and CCONJ bracis-33583 161 17 r2 r2 PROPN bracis-33583 161 18 are be AUX bracis-33583 161 19 the the DET bracis-33583 161 20 lower low ADJ bracis-33583 161 21 and and CCONJ bracis-33583 161 22 upper upper ADJ bracis-33583 161 23 values value NOUN bracis-33583 161 24 of of ADP bracis-33583 161 25 the the DET bracis-33583 161 26 range range NOUN bracis-33583 161 27 of of ADP bracis-33583 161 28 the the DET bracis-33583 161 29 weight weight NOUN bracis-33583 161 30 vector vector NOUN bracis-33583 161 31 distribution distribution NOUN bracis-33583 161 32 . . PUNCT bracis-33583 162 1 fig fig NOUN bracis-33583 162 2 . . PUNCT bracis-33583 163 1 4.adapted 4.adapted NUM bracis-33583 163 2 from from ADP bracis-33583 163 3 alahakoon alahakoon PROPN bracis-33583 163 4 et et PROPN bracis-33583 163 5 al al PROPN bracis-33583 163 6 . . PUNCT bracis-33583 163 7     SPACE bracis-33583 164 1 [ [ X bracis-33583 164 2 6 6 NUM bracis-33583 164 3 ] ] PUNCT bracis-33583 164 4 initializing initialize VERB bracis-33583 164 5 weights weight NOUN bracis-33583 164 6 of of ADP bracis-33583 164 7 a a DET bracis-33583 164 8 new new ADJ bracis-33583 164 9 neuron neuron NOUN bracis-33583 164 10 in in ADP bracis-33583 164 11 a a DET bracis-33583 164 12 grid grid NOUN bracis-33583 164 13 . . PUNCT bracis-33583 165 1 full full ADJ bracis-33583 165 2 size size NOUN bracis-33583 165 3 image image NOUN bracis-33583 165 4 4.4 4.4 NUM bracis-33583 165 5 smoothing smoothing NOUN bracis-33583 165 6 phase phase NOUN bracis-33583 165 7 this this DET bracis-33583 165 8 phase phase NOUN bracis-33583 165 9 adjusts adjust VERB bracis-33583 165 10 the the DET bracis-33583 165 11 neuron neuron NOUN bracis-33583 165 12 ’s ’s PART bracis-33583 165 13 weights weight NOUN bracis-33583 165 14 , , PUNCT bracis-33583 165 15 tuning tune VERB bracis-33583 165 16 the the DET bracis-33583 165 17 grid grid NOUN bracis-33583 165 18 smoother smooth ADJ bracis-33583 165 19 . . PUNCT bracis-33583 166 1 at at ADP bracis-33583 166 2 each each DET bracis-33583 166 3 epoch epoch NOUN bracis-33583 166 4 , , PUNCT bracis-33583 166 5 the the DET bracis-33583 166 6 learning learning NOUN bracis-33583 166 7 rate rate NOUN bracis-33583 166 8 \(\alpha \(\alpha NOUN bracis-33583 166 9 \ \ NOUN bracis-33583 166 10 ) ) PUNCT bracis-33583 166 11 is be AUX bracis-33583 166 12 decreased decrease VERB bracis-33583 166 13 according accord VERB bracis-33583 166 14 to to ADP bracis-33583 166 15 eq eq PROPN bracis-33583 166 16 . . PUNCT bracis-33583 166 17     SPACE bracis-33583 167 1 7 7 NUM bracis-33583 167 2 . . NOUN bracis-33583 167 3 $ $ SYM bracis-33583 167 4 $ $ SYM bracis-33583 167 5 \begin{aligned \begin{aligne VERB bracis-33583 167 6 } } PUNCT bracis-33583 167 7 \alpha \alpha ADP bracis-33583 167 8 ( ( PUNCT bracis-33583 167 9 e+1 e+1 NUM bracis-33583 167 10 ) ) PUNCT bracis-33583 167 11 = = PUNCT bracis-33583 168 1 \alpha \alpha ADP bracis-33583 168 2 \times \time NOUN bracis-33583 168 3 v v NOUN bracis-33583 168 4 ( ( PUNCT bracis-33583 168 5 e e NOUN bracis-33583 168 6 ) ) PUNCT bracis-33583 168 7 \times \time VERB bracis-33583 168 8 \alpha \alpha ADP bracis-33583 168 9 ( ( PUNCT bracis-33583 168 10 e e NOUN bracis-33583 168 11 ) ) PUNCT bracis-33583 168 12 \end{aligned}$$ \end{aligned}$$ X bracis-33583 168 13 ( ( PUNCT bracis-33583 168 14 7 7 NUM bracis-33583 168 15 ) ) PUNCT bracis-33583 168 16 in in ADP bracis-33583 168 17 alahakoon alahakoon NOUN bracis-33583 168 18 et et PROPN bracis-33583 168 19 al al PROPN bracis-33583 168 20 . . PUNCT bracis-33583 168 21     SPACE bracis-33583 169 1 [ [ X bracis-33583 169 2 6 6 NUM bracis-33583 169 3 ] ] PUNCT bracis-33583 169 4 , , PUNCT bracis-33583 169 5 v v NOUN bracis-33583 169 6 is be AUX bracis-33583 169 7 defined define VERB bracis-33583 169 8 as as SCONJ bracis-33583 169 9 shown show VERB bracis-33583 169 10 in in ADP bracis-33583 169 11 eq eq PROPN bracis-33583 169 12 . . PUNCT bracis-33583 169 13     SPACE bracis-33583 169 14 8 8 NUM bracis-33583 169 15 , , PUNCT bracis-33583 169 16 where where SCONJ bracis-33583 169 17 r r NOUN bracis-33583 169 18 is be AUX bracis-33583 169 19 experimentally experimentally ADV bracis-33583 169 20 set set VERB bracis-33583 169 21 to to ADP bracis-33583 169 22 3.8 3.8 NUM bracis-33583 169 23 in in ADP bracis-33583 169 24 cases case NOUN bracis-33583 169 25 of of ADP bracis-33583 169 26 a a DET bracis-33583 169 27 four four NUM bracis-33583 169 28 nodes node NOUN bracis-33583 169 29 initialization initialization NOUN bracis-33583 169 30 grid grid NOUN bracis-33583 169 31 , , PUNCT bracis-33583 169 32 and and CCONJ bracis-33583 169 33 n(e n(e NOUN bracis-33583 169 34 ) ) PUNCT bracis-33583 169 35 the the DET bracis-33583 169 36 number number NOUN bracis-33583 169 37 of of ADP bracis-33583 169 38 nodes node NOUN bracis-33583 169 39 in in ADP bracis-33583 169 40 the the DET bracis-33583 169 41 beginning beginning NOUN bracis-33583 169 42 of of ADP bracis-33583 169 43 that that DET bracis-33583 169 44 epoch epoch NOUN bracis-33583 169 45 . . PUNCT bracis-33583 170 1 $ $ SYM bracis-33583 170 2 $ $ SYM bracis-33583 170 3 \begin{aligned \begin{aligne VERB bracis-33583 170 4 } } PUNCT bracis-33583 170 5 v v NOUN bracis-33583 170 6 ( ( PUNCT bracis-33583 170 7 e e NOUN bracis-33583 170 8 ) ) PUNCT bracis-33583 170 9 = = SYM bracis-33583 170 10 1 1 NUM bracis-33583 170 11 \frac{r}{n(e \frac{r}{n(e NOUN bracis-33583 170 12 ) ) PUNCT bracis-33583 170 13 } } PUNCT bracis-33583 170 14 \end{aligned}$$ \end{aligned}$$ X bracis-33583 170 15 ( ( PUNCT bracis-33583 170 16 8) 8) NUM bracis-33583 170 17 the the DET bracis-33583 170 18 neighborhood neighborhood NOUN bracis-33583 170 19 around around ADP bracis-33583 170 20 the the DET bracis-33583 170 21 winning win VERB bracis-33583 170 22 neuron neuron NOUN bracis-33583 170 23 must must AUX bracis-33583 170 24 also also ADV bracis-33583 170 25 be be AUX bracis-33583 170 26 reduced reduce VERB bracis-33583 170 27 with with ADP bracis-33583 170 28 training training NOUN bracis-33583 170 29 time time NOUN bracis-33583 170 30 . . PUNCT bracis-33583 171 1 this this PRON bracis-33583 171 2 is be AUX bracis-33583 171 3 guaranteed guarantee VERB bracis-33583 171 4 decreasing decrease VERB bracis-33583 171 5 the the DET bracis-33583 171 6 neighborhood neighborhood NOUN bracis-33583 171 7 width width NOUN bracis-33583 171 8 according accord VERB bracis-33583 171 9 to to ADP bracis-33583 171 10 eq eq PROPN bracis-33583 171 11 . . PUNCT bracis-33583 171 12     SPACE bracis-33583 171 13 9 9 NUM bracis-33583 171 14 . . PUNCT bracis-33583 172 1 in in ADP bracis-33583 172 2 the the DET bracis-33583 172 3 equation equation NOUN bracis-33583 172 4 , , PUNCT bracis-33583 172 5 \(\sigma \(\sigma X bracis-33583 172 6 _ _ PUNCT bracis-33583 172 7 0\ 0\ NUM bracis-33583 172 8 ) ) PUNCT bracis-33583 172 9 is be AUX bracis-33583 172 10 the the DET bracis-33583 172 11 neighborhood neighborhood NOUN bracis-33583 172 12 width width NOUN bracis-33583 172 13 in in ADP bracis-33583 172 14 the the DET bracis-33583 172 15 algorithm algorithm NOUN bracis-33583 172 16 ’s ’s PART bracis-33583 172 17 initialization initialization NOUN bracis-33583 172 18 , , PUNCT bracis-33583 172 19 while while SCONJ bracis-33583 172 20 \(\tau \(\tau ADV bracis-33583 172 21 \ \ NOUN bracis-33583 172 22 ) ) PUNCT bracis-33583 172 23 is be AUX bracis-33583 172 24 a a DET bracis-33583 172 25 time time NOUN bracis-33583 172 26 constant constant ADJ bracis-33583 172 27 . . PUNCT bracis-33583 173 1 $ $ SYM bracis-33583 173 2 $ $ SYM bracis-33583 173 3 \begin{aligned \begin{aligne VERB bracis-33583 173 4 } } PUNCT bracis-33583 173 5 \sigma \sigma PROPN bracis-33583 173 6 ( ( PUNCT bracis-33583 173 7 t t NOUN bracis-33583 173 8 ) ) PUNCT bracis-33583 173 9 = = PUNCT bracis-33583 174 1 \sigma \sigma PROPN bracis-33583 174 2 _ _ NOUN bracis-33583 174 3 0 0 PUNCT bracis-33583 175 1 \exp \exp PROPN bracis-33583 175 2 \biggl \biggl PROPN bracis-33583 175 3 ( ( PUNCT bracis-33583 175 4 -\frac{t}{\tau -\frac{t}{\tau SYM bracis-33583 175 5 } } PUNCT bracis-33583 175 6 \biggl \biggl PROPN bracis-33583 175 7 ) ) PUNCT bracis-33583 175 8 \end{aligned}$$ \end{aligned}$$ PROPN bracis-33583 175 9 ( ( PUNCT bracis-33583 175 10 9 9 NUM bracis-33583 175 11 ) ) PUNCT bracis-33583 175 12 4.5 4.5 NUM bracis-33583 175 13 predicting predict VERB bracis-33583 175 14 new new ADJ bracis-33583 175 15 data datum NOUN bracis-33583 175 16 after after ADP bracis-33583 175 17 finishing finish VERB bracis-33583 175 18 training training NOUN bracis-33583 175 19 , , PUNCT bracis-33583 175 20 each each DET bracis-33583 175 21 neuron neuron NOUN bracis-33583 175 22 is be AUX bracis-33583 175 23 associated associate VERB bracis-33583 175 24 with with ADP bracis-33583 175 25 a a DET bracis-33583 175 26 prototype prototype NOUN bracis-33583 175 27 vector vector NOUN bracis-33583 175 28 \(\overline{\boldsymbol{v}}_i\ \(\overline{\boldsymbol{v}}_i\ NOUN bracis-33583 175 29 ) ) PUNCT bracis-33583 175 30 calculated calculate VERB bracis-33583 175 31 the the DET bracis-33583 175 32 same same ADJ bracis-33583 175 33 way way NOUN bracis-33583 175 34 as as SCONJ bracis-33583 175 35 som som NOUN bracis-33583 175 36 - - PUNCT bracis-33583 175 37 mll mll PROPN bracis-33583 175 38 using use VERB bracis-33583 175 39 eq eq PROPN bracis-33583 176 1 . . PUNCT bracis-33583 176 2     SPACE bracis-33583 176 3 4 4 NUM bracis-33583 176 4 . . PUNCT bracis-33583 177 1 a a DET bracis-33583 177 2 threshold threshold NOUN bracis-33583 177 3 is be AUX bracis-33583 177 4 then then ADV bracis-33583 177 5 applied apply VERB bracis-33583 177 6 to to PART bracis-33583 177 7 obtain obtain VERB bracis-33583 177 8 a a DET bracis-33583 177 9 binary binary ADJ bracis-33583 177 10 final final ADJ bracis-33583 177 11 prediction prediction NOUN bracis-33583 177 12 vector vector NOUN bracis-33583 177 13 ( ( PUNCT bracis-33583 177 14 fig fig NOUN bracis-33583 177 15 . . PUNCT bracis-33583 177 16     SPACE bracis-33583 177 17 2 2 NUM bracis-33583 177 18 ) ) PUNCT bracis-33583 177 19 . . PUNCT bracis-33583 178 1 5 5 NUM bracis-33583 178 2 materials material NOUN bracis-33583 178 3 and and CCONJ bracis-33583 178 4     SPACE bracis-33583 178 5 methods method NOUN bracis-33583 178 6 this this DET bracis-33583 178 7 section section NOUN bracis-33583 178 8 presents present VERB bracis-33583 178 9 the the DET bracis-33583 178 10 datasets dataset NOUN bracis-33583 178 11 used use VERB bracis-33583 178 12 in in ADP bracis-33583 178 13 the the DET bracis-33583 178 14 experiments experiment NOUN bracis-33583 178 15 , , PUNCT bracis-33583 178 16 the the DET bracis-33583 178 17 baseline baseline NOUN bracis-33583 178 18 algorithms algorithm NOUN bracis-33583 178 19 used use VERB bracis-33583 178 20 for for ADP bracis-33583 178 21 comparison comparison NOUN bracis-33583 178 22 , , PUNCT bracis-33583 178 23 the the DET bracis-33583 178 24 evaluation evaluation NOUN bracis-33583 178 25 measures measure NOUN bracis-33583 178 26 employed employ VERB bracis-33583 178 27 , , PUNCT bracis-33583 178 28 and and CCONJ bracis-33583 178 29 the the DET bracis-33583 178 30 tuning tuning NOUN bracis-33583 178 31 procedure procedure NOUN bracis-33583 178 32 employed employ VERB bracis-33583 178 33 to to PART bracis-33583 178 34 define define VERB bracis-33583 178 35 the the DET bracis-33583 178 36 hyperparameters hyperparameter NOUN bracis-33583 178 37 of of ADP bracis-33583 178 38 gsom gsom PROPN bracis-33583 178 39 - - PUNCT bracis-33583 178 40 mll mll PROPN bracis-33583 178 41 . . PROPN bracis-33583 179 1 5.1 5.1 NUM bracis-33583 179 2 datasets dataset NOUN bracis-33583 179 3 the the DET bracis-33583 179 4 datasets dataset NOUN bracis-33583 179 5 used use VERB bracis-33583 179 6 are be AUX bracis-33583 179 7 available available ADJ bracis-33583 179 8 in in ADP bracis-33583 179 9 the the DET bracis-33583 179 10 mulan mulan NOUN bracis-33583 179 11 repositoryfootnote repositoryfootnote VERB bracis-33583 179 12 1 1 NUM bracis-33583 179 13 . . PUNCT bracis-33583 179 14 table table NOUN bracis-33583 179 15     SPACE bracis-33583 179 16 1 1 NUM bracis-33583 179 17 shows show VERB bracis-33583 179 18 some some DET bracis-33583 179 19 characteristics characteristic NOUN bracis-33583 179 20 of of ADP bracis-33583 179 21 each each DET bracis-33583 179 22 dataset dataset NOUN bracis-33583 179 23 : : PUNCT bracis-33583 179 24 domain domain NOUN bracis-33583 179 25 , , PUNCT bracis-33583 179 26 the the DET bracis-33583 179 27 number number NOUN bracis-33583 179 28 of of ADP bracis-33583 179 29 instances instance NOUN bracis-33583 179 30 , , PUNCT bracis-33583 179 31 the the DET bracis-33583 179 32 number number NOUN bracis-33583 179 33 of of ADP bracis-33583 179 34 labels label NOUN bracis-33583 179 35 and and CCONJ bracis-33583 179 36 distinct distinct ADJ bracis-33583 179 37 combinations combination NOUN bracis-33583 179 38 of of ADP bracis-33583 179 39 labels label NOUN bracis-33583 179 40 , , PUNCT bracis-33583 179 41 and and CCONJ bracis-33583 179 42 the the DET bracis-33583 179 43 label label NOUN bracis-33583 179 44 cardinality cardinality NOUN bracis-33583 179 45 and and CCONJ bracis-33583 179 46 label label NOUN bracis-33583 179 47 density density NOUN bracis-33583 179 48 . . PUNCT bracis-33583 180 1 label label NOUN bracis-33583 180 2 cardinality cardinality PROPN bracis-33583 180 3 ( ( PUNCT bracis-33583 180 4 lc lc NOUN bracis-33583 180 5 ) ) PUNCT bracis-33583 180 6 is be AUX bracis-33583 180 7 a a DET bracis-33583 180 8 metric metric NOUN bracis-33583 180 9 of of ADP bracis-33583 180 10 the the DET bracis-33583 180 11 average average ADJ bracis-33583 180 12 number number NOUN bracis-33583 180 13 of of ADP bracis-33583 180 14 labels label NOUN bracis-33583 180 15 per per ADP bracis-33583 180 16 instance instance NOUN bracis-33583 180 17 , , PUNCT bracis-33583 180 18 while while SCONJ bracis-33583 180 19 label label NOUN bracis-33583 180 20 density density NOUN bracis-33583 180 21 ( ( PUNCT bracis-33583 180 22 ld ld PROPN bracis-33583 180 23 ) ) PUNCT bracis-33583 180 24 is be AUX bracis-33583 180 25 the the DET bracis-33583 180 26 lc lc PROPN bracis-33583 180 27 metric metric ADJ bracis-33583 180 28 over over ADP bracis-33583 180 29 the the DET bracis-33583 180 30 total total ADJ bracis-33583 180 31 number number NOUN bracis-33583 180 32 of of ADP bracis-33583 180 33 labels label NOUN bracis-33583 180 34 . . PUNCT bracis-33583 181 1 equations equation NOUN bracis-33583 181 2     SPACE bracis-33583 181 3 ( ( PUNCT bracis-33583 181 4 10 10 NUM bracis-33583 181 5 ) ) PUNCT bracis-33583 181 6 and and CCONJ bracis-33583 181 7 ( ( PUNCT bracis-33583 181 8 11 11 X bracis-33583 181 9 ) ) PUNCT bracis-33583 181 10 show show VERB bracis-33583 181 11 ld ld NOUN bracis-33583 181 12 and and CCONJ bracis-33583 181 13 lc lc PROPN bracis-33583 181 14 , , PUNCT bracis-33583 181 15 where where SCONJ bracis-33583 181 16 \(y_i\ \(y_i\ NOUN bracis-33583 181 17 ) ) PUNCT bracis-33583 181 18 is be AUX bracis-33583 181 19 the the DET bracis-33583 181 20 set set NOUN bracis-33583 181 21 of of ADP bracis-33583 181 22 positive positive ADJ bracis-33583 181 23 labels label NOUN bracis-33583 181 24 for for ADP bracis-33583 181 25 instance instance NOUN bracis-33583 181 26 i i PRON bracis-33583 181 27 , , PUNCT bracis-33583 181 28 m m VERB bracis-33583 181 29 is be AUX bracis-33583 181 30 the the DET bracis-33583 181 31 total total ADJ bracis-33583 181 32 number number NOUN bracis-33583 181 33 of of ADP bracis-33583 181 34 instances instance NOUN bracis-33583 181 35 , , PUNCT bracis-33583 181 36 and and CCONJ bracis-33583 181 37 q q NOUN bracis-33583 181 38 is be AUX bracis-33583 181 39 the the DET bracis-33583 181 40 total total ADJ bracis-33583 181 41 number number NOUN bracis-33583 181 42 of of ADP bracis-33583 181 43 labels label NOUN bracis-33583 181 44 . . PUNCT bracis-33583 182 1 table table NOUN bracis-33583 182 2 1 1 NUM bracis-33583 182 3 . . PUNCT bracis-33583 183 1 characteristics characteristic NOUN bracis-33583 183 2 of of ADP bracis-33583 183 3 the the DET bracis-33583 183 4 datasetsfull datasetsfull NOUN bracis-33583 183 5 size size NOUN bracis-33583 183 6 table table NOUN bracis-33583 183 7 $ $ SYM bracis-33583 183 8 $ $ SYM bracis-33583 183 9 \begin{aligned \begin{aligne VERB bracis-33583 183 10 } } PUNCT bracis-33583 183 11 lc lc NOUN bracis-33583 183 12 = = PUNCT bracis-33583 183 13 \frac{1}{m \frac{1}{m VERB bracis-33583 183 14 } } PUNCT bracis-33583 183 15 \sum \sum NOUN bracis-33583 183 16 _ _ PUNCT bracis-33583 183 17 { { PUNCT bracis-33583 183 18 i=1}^{m i=1}^{m PROPN bracis-33583 183 19 } } PUNCT bracis-33583 183 20 { { PUNCT bracis-33583 183 21 |y_i| |y_i| PROPN bracis-33583 183 22 } } PUNCT bracis-33583 183 23 \end{aligned}$$ \end{aligned}$$ X bracis-33583 183 24 ( ( PUNCT bracis-33583 183 25 10 10 NUM bracis-33583 183 26 ) ) PUNCT bracis-33583 183 27 $ $ SYM bracis-33583 183 28 $ $ SYM bracis-33583 183 29 \begin{aligned \begin{aligne VERB bracis-33583 183 30 } } PUNCT bracis-33583 183 31 ld ld NOUN bracis-33583 183 32 = = PUNCT bracis-33583 183 33 \frac{1}{m \frac{1}{m PROPN bracis-33583 183 34 } } PUNCT bracis-33583 183 35 \sum \sum NOUN bracis-33583 183 36 _ _ PUNCT bracis-33583 183 37 { { PUNCT bracis-33583 183 38 i=1}^{m i=1}^{m PROPN bracis-33583 183 39 } } PUNCT bracis-33583 183 40 \frac{|y_i|}{q \frac{|y_i|}{q PROPN bracis-33583 183 41 } } PUNCT bracis-33583 183 42 \end{aligned}$$ \end{aligned}$$ X bracis-33583 183 43 ( ( PUNCT bracis-33583 183 44 11 11 NUM bracis-33583 183 45 ) ) PUNCT bracis-33583 183 46 5.2 5.2 NUM bracis-33583 183 47 baseline baseline NOUN bracis-33583 183 48 classifiers classifier NOUN bracis-33583 183 49 the the DET bracis-33583 183 50 proposed propose VERB bracis-33583 183 51 gsom gsom PROPN bracis-33583 183 52 - - PUNCT bracis-33583 183 53 mll mll PROPN bracis-33583 183 54 was be AUX bracis-33583 183 55 compared compare VERB bracis-33583 183 56 with with ADP bracis-33583 183 57 the the DET bracis-33583 183 58 original original ADJ bracis-33583 183 59 som som PROPN bracis-33583 183 60 - - PUNCT bracis-33583 183 61 mll mll PROPN bracis-33583 183 62 proposed propose VERB bracis-33583 183 63 by by ADP bracis-33583 183 64 colombini colombini PROPN bracis-33583 183 65 et et PROPN bracis-33583 183 66 al al PROPN bracis-33583 183 67 . . PUNCT bracis-33583 183 68     SPACE bracis-33583 184 1 [ [ X bracis-33583 184 2 17 17 NUM bracis-33583 184 3 ] ] PUNCT bracis-33583 184 4 , , PUNCT bracis-33583 184 5 where where SCONJ bracis-33583 184 6 a a DET bracis-33583 184 7 hexagonal hexagonal ADJ bracis-33583 184 8 grid grid NOUN bracis-33583 184 9 of of ADP bracis-33583 184 10 25 25 NUM bracis-33583 184 11 neurons neuron NOUN bracis-33583 184 12 was be AUX bracis-33583 184 13 used use VERB bracis-33583 184 14 , , PUNCT bracis-33583 184 15 the the DET bracis-33583 184 16 learning learning NOUN bracis-33583 184 17 rate rate NOUN bracis-33583 184 18 was be AUX bracis-33583 184 19 set set VERB bracis-33583 184 20 to to PART bracis-33583 184 21 decrease decrease VERB bracis-33583 184 22 linearly linearly ADV bracis-33583 184 23 from from ADP bracis-33583 184 24 0.05 0.05 NUM bracis-33583 184 25 to to ADP bracis-33583 184 26 0.01 0.01 NUM bracis-33583 184 27 at at ADP bracis-33583 184 28 each each DET bracis-33583 184 29 epoch epoch NOUN bracis-33583 184 30 , , PUNCT bracis-33583 184 31 and and CCONJ bracis-33583 184 32 the the DET bracis-33583 184 33 gaussian gaussian ADJ bracis-33583 184 34 function function NOUN bracis-33583 184 35 ( ( PUNCT bracis-33583 184 36 eq eq NOUN bracis-33583 184 37 . . PUNCT bracis-33583 184 38     SPACE bracis-33583 184 39 3 3 X bracis-33583 184 40 ) ) PUNCT bracis-33583 184 41 was be AUX bracis-33583 184 42 used use VERB bracis-33583 184 43 as as ADP bracis-33583 184 44 the the DET bracis-33583 184 45 neighborhood neighborhood NOUN bracis-33583 184 46 with with ADP bracis-33583 184 47 \(\sigma \(\sigma NOUN bracis-33583 184 48 = = SYM bracis-33583 184 49 2/3\ 2/3\ NUM bracis-33583 184 50 ) ) PUNCT bracis-33583 184 51 . . PUNCT bracis-33583 185 1 gsom gsom PROPN bracis-33583 185 2 - - PUNCT bracis-33583 185 3 mll mll PROPN bracis-33583 185 4 was be AUX bracis-33583 185 5 also also ADV bracis-33583 185 6 compared compare VERB bracis-33583 185 7 with with ADP bracis-33583 185 8 different different ADJ bracis-33583 185 9 local local ADJ bracis-33583 185 10 and and CCONJ bracis-33583 185 11 global global ADJ bracis-33583 185 12 - - PUNCT bracis-33583 185 13 based base VERB bracis-33583 185 14 methods method NOUN bracis-33583 185 15 . . PUNCT bracis-33583 186 1 they they PRON bracis-33583 186 2 are be AUX bracis-33583 186 3 listed list VERB bracis-33583 186 4 below below ADV bracis-33583 186 5 . . PUNCT bracis-33583 187 1 binary binary ADJ bracis-33583 187 2 relevance relevance NOUN bracis-33583 187 3 ( ( PUNCT bracis-33583 187 4 br br NOUN bracis-33583 187 5 ) ) PUNCT bracis-33583 187 6 and and CCONJ bracis-33583 187 7 label label NOUN bracis-33583 187 8 powerset powerset NOUN bracis-33583 187 9 ( ( PUNCT bracis-33583 187 10 lp lp NOUN bracis-33583 187 11 ) ) PUNCT bracis-33583 187 12     SPACE bracis-33583 188 1 [ [ X bracis-33583 188 2 35 35 NUM bracis-33583 188 3 ] ] PUNCT bracis-33583 188 4 , , PUNCT bracis-33583 188 5 with with ADP bracis-33583 188 6 the the DET bracis-33583 188 7 following follow VERB bracis-33583 188 8 algorithms algorithm NOUN bracis-33583 188 9 as as ADP bracis-33583 188 10 base base NOUN bracis-33583 188 11 classifiers classifier NOUN bracis-33583 188 12 : : PUNCT bracis-33583 188 13 support support NOUN bracis-33583 188 14 vector vector NOUN bracis-33583 188 15 machine machine NOUN bracis-33583 188 16 ( ( PUNCT bracis-33583 188 17 svm svm PROPN bracis-33583 188 18 ) ) PUNCT bracis-33583 188 19     SPACE bracis-33583 189 1 [ [ X bracis-33583 189 2 38 38 NUM bracis-33583 189 3 ] ] PUNCT bracis-33583 189 4 , , PUNCT bracis-33583 189 5 j48 j48 ADJ bracis-33583 189 6 decision decision NOUN bracis-33583 189 7 tree tree NOUN bracis-33583 189 8 induction induction NOUN bracis-33583 189 9     SPACE bracis-33583 190 1 [ [ X bracis-33583 190 2 24 24 NUM bracis-33583 190 3 ] ] PUNCT bracis-33583 190 4 and and CCONJ bracis-33583 190 5 k k ADV bracis-33583 190 6 - - PUNCT bracis-33583 190 7 nearest near ADJ bracis-33583 190 8 neighbors neighbor NOUN bracis-33583 190 9 ( ( PUNCT bracis-33583 190 10 knn knn PROPN bracis-33583 190 11 ) ) PUNCT bracis-33583 190 12     SPACE bracis-33583 191 1 [ [ X bracis-33583 191 2 5 5 NUM bracis-33583 191 3 ] ] PUNCT bracis-33583 191 4 ; ; PUNCT bracis-33583 191 5 random random ADJ bracis-33583 191 6 k k NOUN bracis-33583 191 7 - - PUNCT bracis-33583 191 8 labelsets labelset NOUN bracis-33583 191 9 ( ( PUNCT bracis-33583 191 10 rakel rakel NOUN bracis-33583 191 11 ) ) PUNCT bracis-33583 191 12     SPACE bracis-33583 192 1 [ [ X bracis-33583 192 2 37 37 NUM bracis-33583 192 3 ] ] PUNCT bracis-33583 192 4 and and CCONJ bracis-33583 192 5 classifiers classifier NOUN bracis-33583 192 6 chains chain NOUN bracis-33583 192 7 ( ( PUNCT bracis-33583 192 8 cc cc NOUN bracis-33583 192 9 ) ) PUNCT bracis-33583 192 10     SPACE bracis-33583 193 1 [ [ X bracis-33583 193 2 27 27 NUM bracis-33583 193 3 ] ] PUNCT bracis-33583 193 4 , , PUNCT bracis-33583 193 5 both both PRON bracis-33583 193 6 also also ADV bracis-33583 193 7 using use VERB bracis-33583 193 8 svm svm PROPN bracis-33583 193 9 , , PUNCT bracis-33583 193 10 j48 j48 PROPN bracis-33583 193 11 and and CCONJ bracis-33583 193 12 knn knn PROPN bracis-33583 193 13 as as ADP bracis-33583 193 14 base base NOUN bracis-33583 193 15 classifiers classifier NOUN bracis-33583 193 16 ; ; PUNCT bracis-33583 193 17 hierarchy hierarchy NOUN bracis-33583 193 18 of of ADP bracis-33583 193 19 multi multi ADJ bracis-33583 193 20 - - ADJ bracis-33583 193 21 label label ADJ bracis-33583 193 22 classifiers classifier NOUN bracis-33583 193 23 ( ( PUNCT bracis-33583 193 24 homer homer NOUN bracis-33583 193 25 ) ) PUNCT bracis-33583 193 26     SPACE bracis-33583 194 1 [ [ X bracis-33583 194 2 23 23 NUM bracis-33583 194 3 ] ] PUNCT bracis-33583 194 4 , , PUNCT bracis-33583 194 5 using use VERB bracis-33583 194 6 lp lp NOUN bracis-33583 194 7 and and CCONJ bracis-33583 194 8 br br NOUN bracis-33583 194 9 , , PUNCT bracis-33583 194 10 with with ADP bracis-33583 194 11 svm svm PROPN bracis-33583 194 12 , , PUNCT bracis-33583 194 13 j48 j48 PROPN bracis-33583 194 14 and and CCONJ bracis-33583 194 15 knn knn PROPN bracis-33583 194 16 as as ADP bracis-33583 194 17 base base NOUN bracis-33583 194 18 classifiers classifier NOUN bracis-33583 194 19 ; ; PUNCT bracis-33583 194 20 back back ADJ bracis-33583 194 21 - - PUNCT bracis-33583 194 22 propagation propagation NOUN bracis-33583 194 23 multi multi ADJ bracis-33583 194 24 - - ADJ bracis-33583 194 25 label label ADJ bracis-33583 194 26 learning learning NOUN bracis-33583 194 27 ( ( PUNCT bracis-33583 194 28 bpmll bpmll NOUN bracis-33583 194 29 ) ) PUNCT bracis-33583 194 30     SPACE bracis-33583 195 1 [ [ X bracis-33583 195 2 42 42 NUM bracis-33583 195 3 ] ] PUNCT bracis-33583 195 4 , , PUNCT bracis-33583 195 5 multi multi ADJ bracis-33583 195 6 - - ADJ bracis-33583 195 7 label label ADJ bracis-33583 195 8 k k NOUN bracis-33583 195 9 - - PUNCT bracis-33583 195 10 nearest near ADJ bracis-33583 195 11 neighbors neighbor NOUN bracis-33583 195 12 ( ( PUNCT bracis-33583 195 13 mlknn mlknn ADJ bracis-33583 195 14 ) ) PUNCT bracis-33583 195 15     SPACE bracis-33583 196 1 [ [ X bracis-33583 196 2 43 43 NUM bracis-33583 196 3 ] ] PUNCT bracis-33583 196 4 , , PUNCT bracis-33583 196 5 instance instance NOUN bracis-33583 196 6 - - PUNCT bracis-33583 196 7 based base VERB bracis-33583 196 8 learning learning NOUN bracis-33583 196 9 by by ADP bracis-33583 196 10 logistic logistic ADJ bracis-33583 196 11 regression regression NOUN bracis-33583 196 12 ( ( PUNCT bracis-33583 196 13 iblr iblr NOUN bracis-33583 196 14 - - PUNCT bracis-33583 196 15 ml ml NOUN bracis-33583 196 16 and and CCONJ bracis-33583 196 17 iblr iblr NOUN bracis-33583 196 18 - - PUNCT bracis-33583 196 19 ml+ ml+ NOUN bracis-33583 196 20 ) ) PUNCT bracis-33583 196 21     SPACE bracis-33583 197 1 [ [ X bracis-33583 197 2 13 13 NUM bracis-33583 197 3 ] ] PUNCT bracis-33583 197 4 , , PUNCT bracis-33583 197 5 and and CCONJ bracis-33583 197 6 predictive predictive ADJ bracis-33583 197 7 clustering clustering ADJ bracis-33583 197 8 trees tree NOUN bracis-33583 197 9 ( ( PUNCT bracis-33583 197 10 clus clus ADJ bracis-33583 197 11 ) ) PUNCT bracis-33583 197 12     SPACE bracis-33583 198 1 [ [ X bracis-33583 198 2 10 10 NUM bracis-33583 198 3 ] ] PUNCT bracis-33583 198 4 . . PUNCT bracis-33583 199 1 all all DET bracis-33583 199 2 the the DET bracis-33583 199 3 baseline baseline ADJ bracis-33583 199 4 classifiers classifier NOUN bracis-33583 199 5 were be AUX bracis-33583 199 6 executed execute VERB bracis-33583 199 7 within within ADP bracis-33583 199 8 mulan mulan NOUN bracis-33583 199 9 , , PUNCT bracis-33583 199 10 a a DET bracis-33583 199 11 library library NOUN bracis-33583 199 12 for for ADP bracis-33583 199 13 multi multi ADJ bracis-33583 199 14 - - ADJ bracis-33583 199 15 label label ADJ bracis-33583 199 16 learning learning NOUN bracis-33583 199 17     SPACE bracis-33583 200 1 [ [ X bracis-33583 200 2 36 36 NUM bracis-33583 200 3 ] ] PUNCT bracis-33583 200 4 , , PUNCT bracis-33583 200 5 using use VERB bracis-33583 200 6 their their PRON bracis-33583 200 7 default default NOUN bracis-33583 200 8 hyperparameter hyperparameter NOUN bracis-33583 200 9 values value NOUN bracis-33583 200 10 . . PUNCT bracis-33583 201 1 the the DET bracis-33583 201 2 knn knn PROPN bracis-33583 201 3 - - PUNCT bracis-33583 201 4 based base VERB bracis-33583 201 5 methods method NOUN bracis-33583 201 6 used use VERB bracis-33583 201 7 \(k \(k NOUN bracis-33583 201 8 = = SYM bracis-33583 201 9 3\ 3\ NUM bracis-33583 201 10 ) ) PUNCT bracis-33583 201 11 . . PUNCT bracis-33583 202 1 bpmll bpmll NOUN bracis-33583 202 2 was be AUX bracis-33583 202 3 executed execute VERB bracis-33583 202 4 with with ADP bracis-33583 202 5 one one NUM bracis-33583 202 6 hidden hide VERB bracis-33583 202 7 layer layer NOUN bracis-33583 202 8 with with ADP bracis-33583 202 9 its its PRON bracis-33583 202 10 number number NOUN bracis-33583 202 11 of of ADP bracis-33583 202 12 neurons neuron NOUN bracis-33583 202 13 equal equal ADJ bracis-33583 202 14 to to ADP bracis-33583 202 15 80 80 NUM bracis-33583 202 16 % % NOUN bracis-33583 202 17 of of ADP bracis-33583 202 18 the the DET bracis-33583 202 19 number number NOUN bracis-33583 202 20 of of ADP bracis-33583 202 21 input input NOUN bracis-33583 202 22 features feature NOUN bracis-33583 202 23 . . PUNCT bracis-33583 203 1 5.3 5.3 NUM bracis-33583 203 2 evaluation evaluation NOUN bracis-33583 203 3 measures measure NOUN bracis-33583 203 4 we we PRON bracis-33583 203 5 evaluated evaluate VERB bracis-33583 203 6 all all DET bracis-33583 203 7 classifiers classifier NOUN bracis-33583 203 8 using use VERB bracis-33583 203 9 the the DET bracis-33583 203 10 well well ADV bracis-33583 203 11 - - PUNCT bracis-33583 203 12 known know VERB bracis-33583 203 13 f1 f1 NOUN bracis-33583 203 14 - - PUNCT bracis-33583 203 15 score score NOUN bracis-33583 203 16 multi multi ADJ bracis-33583 203 17 - - ADJ bracis-33583 203 18 label label ADJ bracis-33583 203 19 measure measure NOUN bracis-33583 203 20     SPACE bracis-33583 204 1 [ [ X bracis-33583 204 2 20 20 NUM bracis-33583 204 3 ] ] PUNCT bracis-33583 204 4 , , PUNCT bracis-33583 204 5 which which PRON bracis-33583 204 6 is be AUX bracis-33583 204 7 the the DET bracis-33583 204 8 harmonic harmonic ADJ bracis-33583 204 9 mean mean NOUN bracis-33583 204 10 of of ADP bracis-33583 204 11 precision precision NOUN bracis-33583 204 12 and and CCONJ bracis-33583 204 13 recall recall NOUN bracis-33583 204 14 . . PUNCT bracis-33583 205 1 its its PRON bracis-33583 205 2 calculation calculation NOUN bracis-33583 205 3 is be AUX bracis-33583 205 4 presented present VERB bracis-33583 205 5 in in ADP bracis-33583 205 6 eq eq ADP bracis-33583 205 7 . . PUNCT bracis-33583 205 8     SPACE bracis-33583 205 9 12 12 NUM bracis-33583 205 10 , , PUNCT bracis-33583 205 11 with with ADP bracis-33583 205 12 \(z_i\ \(z_i\ NOUN bracis-33583 205 13 ) ) PUNCT bracis-33583 205 14 the the DET bracis-33583 205 15 set set NOUN bracis-33583 205 16 of of ADP bracis-33583 205 17 predicted predict VERB bracis-33583 205 18 labels label NOUN bracis-33583 205 19 , , PUNCT bracis-33583 205 20 \(y_i\ \(y_i\ NOUN bracis-33583 205 21 ) ) PUNCT bracis-33583 205 22 the the DET bracis-33583 205 23 set set NOUN bracis-33583 205 24 of of ADP bracis-33583 205 25 true true ADJ bracis-33583 205 26 labels label NOUN bracis-33583 205 27 , , PUNCT bracis-33583 205 28 and and CCONJ bracis-33583 205 29 m m VERB bracis-33583 205 30 the the DET bracis-33583 205 31 number number NOUN bracis-33583 205 32 of of ADP bracis-33583 205 33 instances instance NOUN bracis-33583 205 34 . . PUNCT bracis-33583 206 1 $ $ SYM bracis-33583 206 2 $ $ SYM bracis-33583 206 3 \begin{aligned \begin{aligne VERB bracis-33583 206 4 } } PUNCT bracis-33583 206 5 \text \text ADP bracis-33583 206 6 { { PUNCT bracis-33583 206 7 f1 f1 NOUN bracis-33583 206 8 - - PUNCT bracis-33583 206 9 score score NOUN bracis-33583 206 10 } } PUNCT bracis-33583 206 11 = = SYM bracis-33583 206 12 \frac{1}{m \frac{1}{m ADJ bracis-33583 206 13 } } PUNCT bracis-33583 206 14 \sum \sum NOUN bracis-33583 206 15 _ _ PUNCT bracis-33583 206 16 { { PUNCT bracis-33583 206 17 i=1}^m i=1}^m X bracis-33583 206 18 \frac{2 \frac{2 PROPN bracis-33583 206 19 \times \times ADP bracis-33583 206 20 |y_i |y_i NOUN bracis-33583 206 21 \cap \cap PROPN bracis-33583 206 22 z_i| z_i| PROPN bracis-33583 206 23 } } PUNCT bracis-33583 206 24 { { PUNCT bracis-33583 206 25 |y_i| |y_i| PROPN bracis-33583 206 26 + + CCONJ bracis-33583 206 27 |z_i| |z_i| PROPN bracis-33583 206 28 } } PUNCT bracis-33583 206 29 \end{aligned}$$ \end{aligned}$$ X bracis-33583 206 30 ( ( PUNCT bracis-33583 206 31 12 12 NUM bracis-33583 206 32 ) ) PUNCT bracis-33583 206 33 5.4 5.4 NUM bracis-33583 206 34 hyperparameters hyperparameter NOUN bracis-33583 206 35 and and CCONJ bracis-33583 206 36     SPACE bracis-33583 206 37 tuning tune VERB bracis-33583 206 38 there there ADV bracis-33583 206 39 are be AUX bracis-33583 206 40 many many ADJ bracis-33583 206 41 gsom gsom PROPN bracis-33583 206 42 implementations implementation NOUN bracis-33583 206 43 available available ADJ bracis-33583 206 44 in in ADP bracis-33583 206 45 the the DET bracis-33583 206 46 literature literature NOUN bracis-33583 206 47 . . PUNCT bracis-33583 207 1 adeu adeu PROPN bracis-33583 207 2 et et PROPN bracis-33583 208 1 al al PROPN bracis-33583 208 2 . . PUNCT bracis-33583 208 3     SPACE bracis-33583 209 1 [ [ X bracis-33583 209 2 4 4 NUM bracis-33583 209 3 ] ] PUNCT bracis-33583 209 4 , , PUNCT bracis-33583 209 5 for for ADP bracis-33583 209 6 example example NOUN bracis-33583 209 7 , , PUNCT bracis-33583 209 8 evaluated evaluate VERB bracis-33583 209 9 three three NUM bracis-33583 209 10 of of ADP bracis-33583 209 11 them they PRON bracis-33583 209 12 : : PUNCT bracis-33583 209 13 pygsom pygsom NOUN bracis-33583 209 14 python python NOUN bracis-33583 209 15 package package NOUN bracis-33583 209 16     SPACE bracis-33583 210 1 [ [ X bracis-33583 210 2 3 3 NUM bracis-33583 210 3 ] ] PUNCT bracis-33583 210 4 , , PUNCT bracis-33583 210 5 gsom gsom PROPN bracis-33583 210 6 python python PROPN bracis-33583 210 7 package package NOUN bracis-33583 210 8     SPACE bracis-33583 211 1 [ [ X bracis-33583 211 2 2 2 NUM bracis-33583 211 3 ] ] PUNCT bracis-33583 211 4 and and CCONJ bracis-33583 211 5 the the DET bracis-33583 211 6 growingsom growingsom NOUN bracis-33583 211 7 r r NOUN bracis-33583 211 8 package package NOUN bracis-33583 211 9     SPACE bracis-33583 212 1 [ [ X bracis-33583 212 2 1 1 NUM bracis-33583 212 3 ] ] PUNCT bracis-33583 212 4 . . PUNCT bracis-33583 213 1 these these PRON bracis-33583 213 2 are be AUX bracis-33583 213 3 all all DET bracis-33583 213 4 implementations implementation NOUN bracis-33583 213 5 based base VERB bracis-33583 213 6 on on ADP bracis-33583 213 7 alahakoon alahakoon PROPN bracis-33583 213 8 et et PROPN bracis-33583 213 9 al al PROPN bracis-33583 213 10 . . PUNCT bracis-33583 213 11     SPACE bracis-33583 214 1 [ [ X bracis-33583 214 2 6 6 NUM bracis-33583 214 3 ] ] PUNCT bracis-33583 214 4 , , PUNCT bracis-33583 214 5 and and CCONJ bracis-33583 214 6 according accord VERB bracis-33583 214 7 to to ADP bracis-33583 214 8 adeu adeu PROPN bracis-33583 214 9 et et PROPN bracis-33583 214 10 al al PROPN bracis-33583 214 11 . . PROPN bracis-33583 214 12 , , PUNCT bracis-33583 214 13 all all PRON bracis-33583 214 14 have have VERB bracis-33583 214 15 different different ADJ bracis-33583 214 16 implementation implementation NOUN bracis-33583 214 17 details detail NOUN bracis-33583 214 18 focusing focus VERB bracis-33583 214 19 on on ADP bracis-33583 214 20 specific specific ADJ bracis-33583 214 21 applications application NOUN bracis-33583 214 22 that that PRON bracis-33583 214 23 can can AUX bracis-33583 214 24 slightly slightly ADV bracis-33583 214 25 change change VERB bracis-33583 214 26 the the DET bracis-33583 214 27 results result NOUN bracis-33583 214 28 . . PUNCT bracis-33583 215 1 because because SCONJ bracis-33583 215 2 of of ADP bracis-33583 215 3 this this PRON bracis-33583 215 4 , , PUNCT bracis-33583 215 5 they they PRON bracis-33583 215 6 chose choose VERB bracis-33583 215 7 to to PART bracis-33583 215 8 implement implement VERB bracis-33583 215 9 their their PRON bracis-33583 215 10 own own ADJ bracis-33583 215 11 packagefootnote packagefootnote NOUN bracis-33583 215 12 2 2 NUM bracis-33583 215 13 focusing focus VERB bracis-33583 215 14 on on ADP bracis-33583 215 15 their their PRON bracis-33583 215 16 specific specific ADJ bracis-33583 215 17 application application NOUN bracis-33583 215 18 . . PUNCT bracis-33583 216 1 in in ADP bracis-33583 216 2 the the DET bracis-33583 216 3 same same ADJ bracis-33583 216 4 direction direction NOUN bracis-33583 216 5 , , PUNCT bracis-33583 216 6 we we PRON bracis-33583 216 7 created create VERB bracis-33583 216 8 a a DET bracis-33583 216 9 new new ADJ bracis-33583 216 10 python python NOUN bracis-33583 216 11 package package NOUN bracis-33583 216 12 to to PART bracis-33583 216 13 meet meet VERB bracis-33583 216 14 our our PRON bracis-33583 216 15 needs need NOUN bracis-33583 216 16 here here ADV bracis-33583 216 17 . . PUNCT bracis-33583 217 1 gsom gsom NOUN bracis-33583 217 2 - - PUNCT bracis-33583 217 3 mllfootnote mllfootnote NOUN bracis-33583 217 4 3 3 NUM bracis-33583 217 5 is be AUX bracis-33583 217 6 an an DET bracis-33583 217 7 adaptation adaptation NOUN bracis-33583 217 8 of of ADP bracis-33583 217 9 adeu adeu NOUN bracis-33583 217 10 et et PROPN bracis-33583 217 11 al al PROPN bracis-33583 217 12 . . PROPN bracis-33583 217 13 for for ADP bracis-33583 217 14 multi multi ADJ bracis-33583 217 15 - - ADJ bracis-33583 217 16 label label ADJ bracis-33583 217 17 classification classification NOUN bracis-33583 217 18 . . PUNCT bracis-33583 218 1 we we PRON bracis-33583 218 2 used use VERB bracis-33583 218 3 the the DET bracis-33583 218 4 area area NOUN bracis-33583 218 5 under under ADP bracis-33583 218 6 the the DET bracis-33583 218 7 precision precision NOUN bracis-33583 218 8 - - PUNCT bracis-33583 218 9 recall recall NOUN bracis-33583 218 10 curve curve NOUN bracis-33583 218 11 ( ( PUNCT bracis-33583 218 12 auprc auprc PROPN bracis-33583 218 13 ) ) PUNCT bracis-33583 218 14 to to PART bracis-33583 218 15 find find VERB bracis-33583 218 16 the the DET bracis-33583 218 17 best good ADJ bracis-33583 218 18 hyperparameter hyperparameter NOUN bracis-33583 218 19 values value NOUN bracis-33583 218 20 for for ADP bracis-33583 218 21 our our PRON bracis-33583 218 22 method method NOUN bracis-33583 218 23 . . PUNCT bracis-33583 219 1 a a DET bracis-33583 219 2 precision precision NOUN bracis-33583 219 3 - - PUNCT bracis-33583 219 4 recall recall NOUN bracis-33583 219 5 curve curve NOUN bracis-33583 219 6 is be AUX bracis-33583 219 7 obtained obtain VERB bracis-33583 219 8 by by ADP bracis-33583 219 9 applying apply VERB bracis-33583 219 10 threshold threshold NOUN bracis-33583 219 11 values value NOUN bracis-33583 219 12 in in ADP bracis-33583 219 13 the the DET bracis-33583 219 14 interval interval NOUN bracis-33583 219 15 [ [ X bracis-33583 219 16 0 0 NUM bracis-33583 219 17 , , PUNCT bracis-33583 219 18     SPACE bracis-33583 219 19 1 1 NUM bracis-33583 219 20 ] ] PUNCT bracis-33583 219 21 to to ADP bracis-33583 219 22 the the DET bracis-33583 219 23 classifier classifier NOUN bracis-33583 219 24 ’s ’s PART bracis-33583 219 25 outputs output NOUN bracis-33583 219 26 . . PUNCT bracis-33583 220 1 after after ADP bracis-33583 220 2 interpolation interpolation NOUN bracis-33583 220 3 all all DET bracis-33583 220 4 the the DET bracis-33583 220 5 precision precision NOUN bracis-33583 220 6 - - PUNCT bracis-33583 220 7 recall recall NOUN bracis-33583 220 8 points point NOUN bracis-33583 220 9 ( ( PUNCT bracis-33583 220 10 pr pr NOUN bracis-33583 220 11 - - PUNCT bracis-33583 220 12 points point NOUN bracis-33583 220 13 ) ) PUNCT bracis-33583 220 14 , , PUNCT bracis-33583 220 15 and and CCONJ bracis-33583 220 16 posterior posterior ADJ bracis-33583 220 17 connecting connect VERB bracis-33583 220 18 them they PRON bracis-33583 220 19 , , PUNCT bracis-33583 220 20 we we PRON bracis-33583 220 21 can can AUX bracis-33583 220 22 calculate calculate VERB bracis-33583 220 23 the the DET bracis-33583 220 24 area area NOUN bracis-33583 220 25 under under ADP bracis-33583 220 26 the the DET bracis-33583 220 27 pr pr NOUN bracis-33583 220 28 - - PUNCT bracis-33583 220 29 curve curve NOUN bracis-33583 220 30     SPACE bracis-33583 221 1 [ [ X bracis-33583 221 2 18 18 NUM bracis-33583 221 3 ] ] PUNCT bracis-33583 221 4 . . PUNCT bracis-33583 222 1 we we PRON bracis-33583 222 2 performed perform VERB bracis-33583 222 3 a a DET bracis-33583 222 4 grid grid NOUN bracis-33583 222 5 search search NOUN bracis-33583 222 6 to to PART bracis-33583 222 7 investigate investigate VERB bracis-33583 222 8 different different ADJ bracis-33583 222 9 ranges range NOUN bracis-33583 222 10 of of ADP bracis-33583 222 11 values value NOUN bracis-33583 222 12 for for ADP bracis-33583 222 13 the the DET bracis-33583 222 14 hyperparameters hyperparameter NOUN bracis-33583 222 15 . . PUNCT bracis-33583 223 1 for for ADP bracis-33583 223 2 the the DET bracis-33583 223 3 initial initial ADJ bracis-33583 223 4 learning learning NOUN bracis-33583 223 5 rate rate NOUN bracis-33583 223 6 , , PUNCT bracis-33583 223 7 we we PRON bracis-33583 223 8 tested test VERB bracis-33583 223 9 20 20 NUM bracis-33583 223 10 values value NOUN bracis-33583 223 11 ranging range VERB bracis-33583 223 12 from from ADP bracis-33583 223 13 0.005 0.005 NUM bracis-33583 223 14 to to ADP bracis-33583 223 15 0.195 0.195 NUM bracis-33583 223 16 ; ; PUNCT bracis-33583 223 17 for for ADP bracis-33583 223 18 the the DET bracis-33583 223 19 spread spread ADJ bracis-33583 223 20 factor factor NOUN bracis-33583 223 21 , , PUNCT bracis-33583 223 22 we we PRON bracis-33583 223 23 evaluated evaluate VERB bracis-33583 223 24 20 20 NUM bracis-33583 223 25 values value NOUN bracis-33583 223 26 ranging range VERB bracis-33583 223 27 from from ADP bracis-33583 223 28 0.05 0.05 NUM bracis-33583 223 29 to to ADP bracis-33583 223 30 0.95 0.95 NUM bracis-33583 223 31 ; ; PUNCT bracis-33583 223 32 and and CCONJ bracis-33583 223 33 for for ADP bracis-33583 223 34 the the DET bracis-33583 223 35 smoothing smoothing NOUN bracis-33583 223 36 and and CCONJ bracis-33583 223 37 growing grow VERB bracis-33583 223 38 factors factor NOUN bracis-33583 223 39 , , PUNCT bracis-33583 223 40 we we PRON bracis-33583 223 41 tested test VERB bracis-33583 223 42 values value NOUN bracis-33583 223 43 ranging range VERB bracis-33583 223 44 from from ADP bracis-33583 223 45 1 1 NUM bracis-33583 223 46 to to ADP bracis-33583 223 47 5 5 NUM bracis-33583 223 48 and and CCONJ bracis-33583 223 49 1 1 NUM bracis-33583 223 50 to to PART bracis-33583 223 51 10 10 NUM bracis-33583 223 52 , , PUNCT bracis-33583 223 53 respectively respectively ADV bracis-33583 223 54 . . PUNCT bracis-33583 224 1 we we PRON bracis-33583 224 2 chose choose VERB bracis-33583 224 3 the the DET bracis-33583 224 4 combination combination NOUN bracis-33583 224 5 of of ADP bracis-33583 224 6 values value NOUN bracis-33583 224 7 that that PRON bracis-33583 224 8 led lead VERB bracis-33583 224 9 to to ADP bracis-33583 224 10 the the DET bracis-33583 224 11 best good ADJ bracis-33583 224 12 auprc auprc ADJ bracis-33583 224 13 value value NOUN bracis-33583 224 14 . . PUNCT bracis-33583 225 1 the the DET bracis-33583 225 2 gaussian gaussian ADJ bracis-33583 225 3 function function NOUN bracis-33583 225 4 with with ADP bracis-33583 225 5 \(\sigma \(\sigma NOUN bracis-33583 225 6 = = SYM bracis-33583 225 7 1\ 1\ NUM bracis-33583 225 8 ) ) PUNCT bracis-33583 225 9 for for ADP bracis-33583 225 10 both both CCONJ bracis-33583 225 11 growing grow VERB bracis-33583 225 12 and and CCONJ bracis-33583 225 13 smoothing smooth VERB bracis-33583 225 14 phases phase NOUN bracis-33583 225 15 was be AUX bracis-33583 225 16 used use VERB bracis-33583 225 17 as as ADP bracis-33583 225 18 the the DET bracis-33583 225 19 neighborhood neighborhood NOUN bracis-33583 225 20 influence influence NOUN bracis-33583 225 21 . . PUNCT bracis-33583 226 1 after after ADP bracis-33583 226 2 obtaining obtain VERB bracis-33583 226 3 the the DET bracis-33583 226 4 best good ADJ bracis-33583 226 5 hyperparameter hyperparameter NOUN bracis-33583 226 6 values value NOUN bracis-33583 226 7 , , PUNCT bracis-33583 226 8 the the DET bracis-33583 226 9 threshold threshold NOUN bracis-33583 226 10 of of ADP bracis-33583 226 11 0.5 0.5 NUM bracis-33583 226 12 was be AUX bracis-33583 226 13 applied apply VERB bracis-33583 226 14 to to ADP bracis-33583 226 15 the the DET bracis-33583 226 16 prototype prototype NOUN bracis-33583 226 17 vectors vector NOUN bracis-33583 226 18 in in ADP bracis-33583 226 19 order order NOUN bracis-33583 226 20 to to PART bracis-33583 226 21 obtain obtain VERB bracis-33583 226 22 the the DET bracis-33583 226 23 final final ADJ bracis-33583 226 24 binary binary ADJ bracis-33583 226 25 predictions prediction NOUN bracis-33583 226 26 . . PUNCT bracis-33583 227 1 all all DET bracis-33583 227 2 grids grid NOUN bracis-33583 227 3 were be AUX bracis-33583 227 4 initialized initialize VERB bracis-33583 227 5 with with ADP bracis-33583 227 6 4 4 NUM bracis-33583 227 7 neurons neuron NOUN bracis-33583 227 8 . . PUNCT bracis-33583 228 1 6 6 NUM bracis-33583 228 2 experiments experiment NOUN bracis-33583 228 3 and and CCONJ bracis-33583 228 4     SPACE bracis-33583 228 5 discussion discussion NOUN bracis-33583 228 6 all all DET bracis-33583 228 7 the the DET bracis-33583 228 8 experiments experiment NOUN bracis-33583 228 9 were be AUX bracis-33583 228 10 performed perform VERB bracis-33583 228 11 using use VERB bracis-33583 228 12 the the DET bracis-33583 228 13 iterative iterative NOUN bracis-33583 228 14 stratified stratify VERB bracis-33583 228 15 10 10 NUM bracis-33583 228 16 - - ADJ bracis-33583 228 17 fold fold ADJ bracis-33583 228 18 cross cross ADJ bracis-33583 228 19 - - ADJ bracis-33583 228 20 validation validation NOUN bracis-33583 228 21 proposed propose VERB bracis-33583 228 22 for for ADP bracis-33583 228 23 multi multi ADJ bracis-33583 228 24 - - ADJ bracis-33583 228 25 label label ADJ bracis-33583 228 26 data datum NOUN bracis-33583 228 27     SPACE bracis-33583 229 1 [ [ X bracis-33583 229 2 31 31 NUM bracis-33583 229 3 ] ] PUNCT bracis-33583 229 4 . . PUNCT bracis-33583 230 1 tables table NOUN bracis-33583 230 2     SPACE bracis-33583 230 3 2 2 NUM bracis-33583 230 4 , , PUNCT bracis-33583 230 5 3 3 NUM bracis-33583 230 6 , , PUNCT bracis-33583 230 7 and and CCONJ bracis-33583 230 8 4 4 NUM bracis-33583 230 9 present present ADJ bracis-33583 230 10 , , PUNCT bracis-33583 230 11 respectively respectively ADV bracis-33583 230 12 , , PUNCT bracis-33583 230 13 the the DET bracis-33583 230 14 average average ADJ bracis-33583 230 15 f1 f1 ADJ bracis-33583 230 16 - - PUNCT bracis-33583 230 17 score score NOUN bracis-33583 230 18 values value NOUN bracis-33583 230 19 and and CCONJ bracis-33583 230 20 standard standard ADJ bracis-33583 230 21 deviations deviation NOUN bracis-33583 230 22 over over ADP bracis-33583 230 23 the the DET bracis-33583 230 24 10 10 NUM bracis-33583 230 25 executions execution NOUN bracis-33583 230 26 for for ADP bracis-33583 230 27 each each DET bracis-33583 230 28 dataset dataset NOUN bracis-33583 230 29 and and CCONJ bracis-33583 230 30 method method NOUN bracis-33583 230 31 . . PUNCT bracis-33583 231 1 the the DET bracis-33583 231 2 best good ADJ bracis-33583 231 3 results result NOUN bracis-33583 231 4 are be AUX bracis-33583 231 5 highlighted highlight VERB bracis-33583 231 6 in in ADP bracis-33583 231 7 boldface boldface NOUN bracis-33583 231 8 . . PUNCT bracis-33583 232 1 table table NOUN bracis-33583 232 2 2 2 NUM bracis-33583 232 3 . . PUNCT bracis-33583 232 4 comparing compare VERB bracis-33583 232 5 gsom gsom PROPN bracis-33583 232 6 - - PUNCT bracis-33583 232 7 mll mll PROPN bracis-33583 232 8 with with ADP bracis-33583 232 9 the the DET bracis-33583 232 10 strategies strategy NOUN bracis-33583 232 11 br br NOUN bracis-33583 232 12 , , PUNCT bracis-33583 232 13 lp lp NOUN bracis-33583 232 14 , , PUNCT bracis-33583 232 15 and and CCONJ bracis-33583 232 16 ccfull ccfull NOUN bracis-33583 232 17 size size NOUN bracis-33583 232 18 table table NOUN bracis-33583 232 19 table table NOUN bracis-33583 232 20 3 3 NUM bracis-33583 232 21 . . PUNCT bracis-33583 232 22 comparing compare VERB bracis-33583 232 23 gsom gsom PROPN bracis-33583 232 24 - - PUNCT bracis-33583 232 25 mll mll PROPN bracis-33583 232 26 with with ADP bracis-33583 232 27 the the DET bracis-33583 232 28 methods method NOUN bracis-33583 232 29 rakel rakel NOUN bracis-33583 232 30 and and CCONJ bracis-33583 232 31 homerfull homerfull ADJ bracis-33583 232 32 size size NOUN bracis-33583 232 33 table table NOUN bracis-33583 232 34 table table NOUN bracis-33583 232 35 4 4 NUM bracis-33583 232 36 . . PUNCT bracis-33583 232 37 comparing compare VERB bracis-33583 232 38 gsom gsom PROPN bracis-33583 232 39 - - PUNCT bracis-33583 232 40 mll mll PROPN bracis-33583 232 41 with with ADP bracis-33583 232 42 som som PROPN bracis-33583 232 43 - - PUNCT bracis-33583 232 44 mll mll PROPN bracis-33583 232 45 , , PUNCT bracis-33583 232 46 mlknn mlknn ADJ bracis-33583 232 47 , , PUNCT bracis-33583 232 48 bpmll bpmll NOUN bracis-33583 232 49 , , PUNCT bracis-33583 232 50 iblr iblr NOUN bracis-33583 232 51 and and CCONJ bracis-33583 232 52 clusfull clusfull ADJ bracis-33583 232 53 size size NOUN bracis-33583 232 54 table table NOUN bracis-33583 232 55 the the DET bracis-33583 232 56 overall overall ADJ bracis-33583 232 57 results result NOUN bracis-33583 232 58 show show VERB bracis-33583 232 59 that that SCONJ bracis-33583 232 60 gsom gsom PROPN bracis-33583 232 61 - - PUNCT bracis-33583 232 62 mll mll PROPN bracis-33583 232 63 is be AUX bracis-33583 232 64 superior superior ADJ bracis-33583 232 65 to to ADP bracis-33583 232 66 som som NOUN bracis-33583 232 67 - - PUNCT bracis-33583 232 68 mll mll PROPN bracis-33583 232 69 and and CCONJ bracis-33583 232 70 competitive competitive ADJ bracis-33583 232 71 with with ADP bracis-33583 232 72 the the DET bracis-33583 232 73 performances performance NOUN bracis-33583 232 74 of of ADP bracis-33583 232 75 the the DET bracis-33583 232 76 other other ADJ bracis-33583 232 77 methods method NOUN bracis-33583 232 78 investigated investigate VERB bracis-33583 232 79 . . PUNCT bracis-33583 233 1 the the DET bracis-33583 233 2 gsom gsom PROPN bracis-33583 233 3 - - PUNCT bracis-33583 233 4 mll mll PROPN bracis-33583 233 5 f1 f1 PROPN bracis-33583 233 6 - - PUNCT bracis-33583 233 7 score score NOUN bracis-33583 233 8 was be AUX bracis-33583 233 9 superior superior ADJ bracis-33583 233 10 to to ADP bracis-33583 233 11 som som NOUN bracis-33583 233 12 - - PUNCT bracis-33583 233 13 mll mll PROPN bracis-33583 233 14 , , PUNCT bracis-33583 233 15 especially especially ADV bracis-33583 233 16 in in ADP bracis-33583 233 17 the the DET bracis-33583 233 18 dataset dataset NOUN bracis-33583 233 19 flags flag NOUN bracis-33583 233 20 and and CCONJ bracis-33583 233 21     SPACE bracis-33583 233 22 scene scene NOUN bracis-33583 233 23 . . PUNCT bracis-33583 234 1 we we PRON bracis-33583 234 2 can can AUX bracis-33583 234 3 observe observe VERB bracis-33583 234 4 that that DET bracis-33583 234 5 gsom gsom PROPN bracis-33583 234 6 - - PUNCT bracis-33583 234 7 mll mll PROPN bracis-33583 234 8 performed perform VERB bracis-33583 234 9 surprisingly surprisingly ADV bracis-33583 234 10 badly badly ADV bracis-33583 234 11 in in ADP bracis-33583 234 12 the the DET bracis-33583 234 13 birds bird NOUN bracis-33583 234 14 dataset dataset VERB bracis-33583 234 15 . . PUNCT bracis-33583 235 1 this this DET bracis-33583 235 2 dataset dataset NOUN bracis-33583 235 3 has have VERB bracis-33583 235 4 the the DET bracis-33583 235 5 smallest small ADJ bracis-33583 235 6 label label NOUN bracis-33583 235 7 density density NOUN bracis-33583 235 8 , , PUNCT bracis-33583 235 9 meaning mean VERB bracis-33583 235 10 it it PRON bracis-33583 235 11 has have VERB bracis-33583 235 12 a a DET bracis-33583 235 13 very very ADV bracis-33583 235 14 sparse sparse ADJ bracis-33583 235 15 distribution distribution NOUN bracis-33583 235 16 of of ADP bracis-33583 235 17 labels label NOUN bracis-33583 235 18 assigned assign VERB bracis-33583 235 19 for for ADP bracis-33583 235 20 each each DET bracis-33583 235 21 dataset dataset NOUN bracis-33583 235 22 instance instance NOUN bracis-33583 235 23 . . PUNCT bracis-33583 236 1 our our PRON bracis-33583 236 2 tuning tuning NOUN bracis-33583 236 3 procedure procedure NOUN bracis-33583 236 4 may may AUX bracis-33583 236 5 not not PART bracis-33583 236 6 have have AUX bracis-33583 236 7 been be AUX bracis-33583 236 8 effective effective ADJ bracis-33583 236 9 in in ADP bracis-33583 236 10 maximizing maximize VERB bracis-33583 236 11 performance performance NOUN bracis-33583 236 12 in in ADP bracis-33583 236 13 these these DET bracis-33583 236 14 circumstances circumstance NOUN bracis-33583 236 15 . . PUNCT bracis-33583 237 1 overall overall ADV bracis-33583 237 2 , , PUNCT bracis-33583 237 3 the the DET bracis-33583 237 4 experiments experiment NOUN bracis-33583 237 5 show show VERB bracis-33583 237 6 us we PRON bracis-33583 237 7 that that PRON bracis-33583 237 8 dynamically dynamically ADV bracis-33583 237 9 increasing increase VERB bracis-33583 237 10 the the DET bracis-33583 237 11 neuron neuron NOUN bracis-33583 237 12 grid grid NOUN bracis-33583 237 13 dimension dimension NOUN bracis-33583 237 14 is be AUX bracis-33583 237 15 better well ADJ bracis-33583 237 16 than than ADP bracis-33583 237 17 working work VERB bracis-33583 237 18 with with ADP bracis-33583 237 19 a a DET bracis-33583 237 20 fixed fix VERB bracis-33583 237 21 grid grid NOUN bracis-33583 237 22 size size NOUN bracis-33583 237 23 . . PUNCT bracis-33583 238 1 although although SCONJ bracis-33583 238 2 the the DET bracis-33583 238 3 f1 f1 NOUN bracis-33583 238 4 - - PUNCT bracis-33583 238 5 score score NOUN bracis-33583 238 6 results result NOUN bracis-33583 238 7 were be AUX bracis-33583 238 8 not not PART bracis-33583 238 9 as as ADV bracis-33583 238 10 satisfactory satisfactory ADJ bracis-33583 238 11 as as SCONJ bracis-33583 238 12 expected expect VERB bracis-33583 238 13 , , PUNCT bracis-33583 238 14 gsom gsom PROPN bracis-33583 238 15 - - PUNCT bracis-33583 238 16 mll mll PROPN bracis-33583 238 17 is be AUX bracis-33583 238 18 competitive competitive ADJ bracis-33583 238 19 with with ADP bracis-33583 238 20 the the DET bracis-33583 238 21 other other ADJ bracis-33583 238 22 methods method NOUN bracis-33583 238 23 and and CCONJ bracis-33583 238 24 is be AUX bracis-33583 238 25 a a DET bracis-33583 238 26 promising promising ADJ bracis-33583 238 27 strategy strategy NOUN bracis-33583 238 28 , , PUNCT bracis-33583 238 29 especially especially ADV bracis-33583 238 30 for for ADP bracis-33583 238 31 data data NOUN bracis-33583 238 32 visualization visualization NOUN bracis-33583 238 33 , , PUNCT bracis-33583 238 34 since since SCONJ bracis-33583 238 35 we we PRON bracis-33583 238 36 can can AUX bracis-33583 238 37 observe observe VERB bracis-33583 238 38 how how SCONJ bracis-33583 238 39 instances instance NOUN bracis-33583 238 40 are be AUX bracis-33583 238 41 grouped group VERB bracis-33583 238 42 in in ADP bracis-33583 238 43 the the DET bracis-33583 238 44 grid grid NOUN bracis-33583 238 45 . . PUNCT bracis-33583 239 1 we we PRON bracis-33583 239 2 can can AUX bracis-33583 239 3 have have VERB bracis-33583 239 4 a a DET bracis-33583 239 5 2d 2d NUM bracis-33583 239 6 visualization visualization NOUN bracis-33583 239 7 , , PUNCT bracis-33583 239 8 keeping keep VERB bracis-33583 239 9 the the DET bracis-33583 239 10 original original ADJ bracis-33583 239 11 properties property NOUN bracis-33583 239 12 of of ADP bracis-33583 239 13 the the DET bracis-33583 239 14 input input NOUN bracis-33583 239 15 space space NOUN bracis-33583 239 16 . . PUNCT bracis-33583 240 1 table table NOUN bracis-33583 240 2     SPACE bracis-33583 240 3 5 5 NUM bracis-33583 240 4 presents present VERB bracis-33583 240 5 the the DET bracis-33583 240 6 mean mean ADJ bracis-33583 240 7 values value NOUN bracis-33583 240 8 for for ADP bracis-33583 240 9 each each DET bracis-33583 240 10 hyperparameter hyperparameter NOUN bracis-33583 240 11 after after ADP bracis-33583 240 12 the the DET bracis-33583 240 13 tuning tuning NOUN bracis-33583 240 14 over over ADP bracis-33583 240 15 the the DET bracis-33583 240 16 10 10 NUM bracis-33583 240 17 - - ADJ bracis-33583 240 18 fold fold ADJ bracis-33583 240 19 cross cross ADJ bracis-33583 240 20 - - ADJ bracis-33583 240 21 validation validation ADJ bracis-33583 240 22 procedure procedure NOUN bracis-33583 240 23 : : PUNCT bracis-33583 240 24 the the DET bracis-33583 240 25 number number NOUN bracis-33583 240 26 of of ADP bracis-33583 240 27 growing grow VERB bracis-33583 240 28 and and CCONJ bracis-33583 240 29 smooth smooth ADJ bracis-33583 240 30 epochs epoch NOUN bracis-33583 240 31 , , PUNCT bracis-33583 240 32 the the DET bracis-33583 240 33 spread spread ADJ bracis-33583 240 34 factors factor NOUN bracis-33583 240 35 , , PUNCT bracis-33583 240 36 the the DET bracis-33583 240 37 learning learning NOUN bracis-33583 240 38 rates rate NOUN bracis-33583 240 39 , , PUNCT bracis-33583 240 40 the the DET bracis-33583 240 41 time time NOUN bracis-33583 240 42 in in ADP bracis-33583 240 43 seconds second NOUN bracis-33583 240 44 to to PART bracis-33583 240 45 run run VERB bracis-33583 240 46 the the DET bracis-33583 240 47 model model NOUN bracis-33583 240 48 , , PUNCT bracis-33583 240 49 and and CCONJ bracis-33583 240 50 the the DET bracis-33583 240 51 auprc auprc ADJ bracis-33583 240 52 values value NOUN bracis-33583 240 53 . . PUNCT bracis-33583 241 1 table table NOUN bracis-33583 241 2 5 5 NUM bracis-33583 241 3 . . PUNCT bracis-33583 242 1 mean mean ADJ bracis-33583 242 2 values value NOUN bracis-33583 242 3 of of ADP bracis-33583 242 4 the the DET bracis-33583 242 5 gsom gsom PROPN bracis-33583 242 6 hyperparameters hyperparameter NOUN bracis-33583 242 7 after after ADP bracis-33583 242 8 tuningfull tuningfull ADJ bracis-33583 242 9 size size NOUN bracis-33583 242 10 table table NOUN bracis-33583 242 11 table table NOUN bracis-33583 242 12     SPACE bracis-33583 242 13 5 5 NUM bracis-33583 242 14 also also ADV bracis-33583 242 15 shows show VERB bracis-33583 242 16 that that SCONJ bracis-33583 242 17 a a DET bracis-33583 242 18 high high ADJ bracis-33583 242 19 spread spread NOUN bracis-33583 242 20 factor factor NOUN bracis-33583 242 21 is be AUX bracis-33583 242 22 preferred prefer VERB bracis-33583 242 23 for for ADP bracis-33583 242 24 most most ADJ bracis-33583 242 25 datasets dataset NOUN bracis-33583 242 26 , , PUNCT bracis-33583 242 27 indicating indicate VERB bracis-33583 242 28 a a DET bracis-33583 242 29 large large ADJ bracis-33583 242 30 grid grid NOUN bracis-33583 242 31 instead instead ADV bracis-33583 242 32 of of ADP bracis-33583 242 33 a a DET bracis-33583 242 34 grid grid NOUN bracis-33583 242 35 with with ADP bracis-33583 242 36 few few ADJ bracis-33583 242 37 neurons neuron NOUN bracis-33583 242 38 . . PUNCT bracis-33583 243 1 another another DET bracis-33583 243 2 observation observation NOUN bracis-33583 243 3 is be AUX bracis-33583 243 4 that that SCONJ bracis-33583 243 5 the the DET bracis-33583 243 6 learning learning NOUN bracis-33583 243 7 rate rate NOUN bracis-33583 243 8 is be AUX bracis-33583 243 9 well well ADV bracis-33583 243 10 - - PUNCT bracis-33583 243 11 suited suit VERB bracis-33583 243 12 to to ADP bracis-33583 243 13 most most ADJ bracis-33583 243 14 datasets dataset NOUN bracis-33583 243 15 , , PUNCT bracis-33583 243 16 except except SCONJ bracis-33583 243 17 for for ADP bracis-33583 243 18 birds bird NOUN bracis-33583 243 19 , , PUNCT bracis-33583 243 20 which which PRON bracis-33583 243 21 is be AUX bracis-33583 243 22 very very ADV bracis-33583 243 23 close close ADJ bracis-33583 243 24 to to ADP bracis-33583 243 25 the the DET bracis-33583 243 26 maximum maximum ADJ bracis-33583 243 27 value value NOUN bracis-33583 243 28 used use VERB bracis-33583 243 29 during during ADP bracis-33583 243 30 the the DET bracis-33583 243 31 tuning tuning NOUN bracis-33583 243 32 stage stage NOUN bracis-33583 243 33 ( ( PUNCT bracis-33583 243 34 0.195 0.195 NUM bracis-33583 243 35 ) ) PUNCT bracis-33583 243 36 . . PUNCT bracis-33583 244 1 table table NOUN bracis-33583 244 2     SPACE bracis-33583 244 3 6 6 NUM bracis-33583 244 4 shows show VERB bracis-33583 244 5 each each DET bracis-33583 244 6 dataset dataset NOUN bracis-33583 244 7 ’s ’s PART bracis-33583 244 8 final final ADJ bracis-33583 244 9 number number NOUN bracis-33583 244 10 of of ADP bracis-33583 244 11 neurons neuron NOUN bracis-33583 244 12 in in ADP bracis-33583 244 13 the the DET bracis-33583 244 14 grid grid NOUN bracis-33583 244 15 . . PUNCT bracis-33583 245 1 the the DET bracis-33583 245 2 results result NOUN bracis-33583 245 3 of of ADP bracis-33583 245 4 the the DET bracis-33583 245 5 gsom gsom PROPN bracis-33583 245 6 - - PUNCT bracis-33583 245 7 mll mll PROPN bracis-33583 245 8 experiment experiment NOUN bracis-33583 245 9 suggest suggest VERB bracis-33583 245 10 that that SCONJ bracis-33583 245 11 the the DET bracis-33583 245 12 number number NOUN bracis-33583 245 13 of of ADP bracis-33583 245 14 neurons neuron NOUN bracis-33583 245 15 in in ADP bracis-33583 245 16 the the DET bracis-33583 245 17 grid grid NOUN bracis-33583 245 18 significantly significantly ADV bracis-33583 245 19 influenced influence VERB bracis-33583 245 20 the the DET bracis-33583 245 21 outcomes outcome NOUN bracis-33583 245 22 . . PUNCT bracis-33583 246 1 this this PRON bracis-33583 246 2 is be AUX bracis-33583 246 3 because because SCONJ bracis-33583 246 4 all all DET bracis-33583 246 5 the the DET bracis-33583 246 6 datasets dataset NOUN bracis-33583 246 7 in in ADP bracis-33583 246 8 the the DET bracis-33583 246 9 original original ADJ bracis-33583 246 10 som som PROPN bracis-33583 246 11 - - PUNCT bracis-33583 246 12 mll mll PROPN bracis-33583 246 13 experiment experiment NOUN bracis-33583 246 14 had have VERB bracis-33583 246 15 a a DET bracis-33583 246 16 fixed fix VERB bracis-33583 246 17 grid grid NOUN bracis-33583 246 18 of of ADP bracis-33583 246 19 25 25 NUM bracis-33583 246 20 neurons neuron NOUN bracis-33583 246 21 , , PUNCT bracis-33583 246 22 while while SCONJ bracis-33583 246 23 in in ADP bracis-33583 246 24 gsom gsom PROPN bracis-33583 246 25 - - PUNCT bracis-33583 246 26 mll mll PROPN bracis-33583 246 27 experiment experiment NOUN bracis-33583 246 28 , , PUNCT bracis-33583 246 29 all all DET bracis-33583 246 30 datasets dataset NOUN bracis-33583 246 31 started start VERB bracis-33583 246 32 with with ADP bracis-33583 246 33 a a DET bracis-33583 246 34 grid grid NOUN bracis-33583 246 35 of of ADP bracis-33583 246 36 4 4 NUM bracis-33583 246 37 neurons neuron NOUN bracis-33583 246 38 . . PUNCT bracis-33583 247 1 table table NOUN bracis-33583 247 2 6 6 NUM bracis-33583 247 3 . . PUNCT bracis-33583 247 4 final final ADJ bracis-33583 247 5 number number NOUN bracis-33583 247 6 of of ADP bracis-33583 247 7 neurons neuron NOUN bracis-33583 247 8 in in ADP bracis-33583 247 9 the the DET bracis-33583 247 10 gsom gsom PROPN bracis-33583 247 11 - - PUNCT bracis-33583 247 12 mll mll PROPN bracis-33583 247 13 gridfull gridfull PROPN bracis-33583 247 14 size size NOUN bracis-33583 247 15 table table NOUN bracis-33583 247 16 another another DET bracis-33583 247 17 point point NOUN bracis-33583 247 18 to to PART bracis-33583 247 19 consider consider VERB bracis-33583 247 20 is be AUX bracis-33583 247 21 that that SCONJ bracis-33583 247 22 the the DET bracis-33583 247 23 time time NOUN bracis-33583 247 24 spent spend VERB bracis-33583 247 25 on on ADP bracis-33583 247 26 each each DET bracis-33583 247 27 iteration iteration NOUN bracis-33583 247 28 of of ADP bracis-33583 247 29 the the DET bracis-33583 247 30 algorithm algorithm NOUN bracis-33583 247 31 correlates correlate VERB bracis-33583 247 32 with with ADP bracis-33583 247 33 the the DET bracis-33583 247 34 number number NOUN bracis-33583 247 35 of of ADP bracis-33583 247 36 neurons neuron NOUN bracis-33583 247 37 on on ADP bracis-33583 247 38 the the DET bracis-33583 247 39 final final ADJ bracis-33583 247 40 grid grid NOUN bracis-33583 247 41 . . PUNCT bracis-33583 248 1 this this PRON bracis-33583 248 2 is be AUX bracis-33583 248 3 because because SCONJ bracis-33583 248 4 more more ADJ bracis-33583 248 5 neurons neuron NOUN bracis-33583 248 6 can can AUX bracis-33583 248 7 lead lead VERB bracis-33583 248 8 to to ADP bracis-33583 248 9 increased increase VERB bracis-33583 248 10 computational computational ADJ bracis-33583 248 11 costs cost NOUN bracis-33583 248 12 . . PUNCT bracis-33583 249 1 7 7 NUM bracis-33583 249 2 conclusion conclusion NOUN bracis-33583 249 3 and and CCONJ bracis-33583 249 4     SPACE bracis-33583 249 5 future future ADJ bracis-33583 249 6 work work NOUN bracis-33583 249 7 in in ADP bracis-33583 249 8 this this DET bracis-33583 249 9 work work NOUN bracis-33583 249 10 , , PUNCT bracis-33583 249 11 we we PRON bracis-33583 249 12 presented present VERB bracis-33583 249 13 gsom gsom PROPN bracis-33583 249 14 - - PUNCT bracis-33583 249 15 mll mll PROPN bracis-33583 249 16 , , PUNCT bracis-33583 249 17 a a DET bracis-33583 249 18 growing grow VERB bracis-33583 249 19 self self NOUN bracis-33583 249 20 - - PUNCT bracis-33583 249 21 organizing organize VERB bracis-33583 249 22 map map NOUN bracis-33583 249 23 tailored tailor VERB bracis-33583 249 24 for for ADP bracis-33583 249 25 multi multi ADJ bracis-33583 249 26 - - ADJ bracis-33583 249 27 label label ADJ bracis-33583 249 28 classification classification NOUN bracis-33583 249 29 . . PUNCT bracis-33583 250 1 the the DET bracis-33583 250 2 model model NOUN bracis-33583 250 3 ’s ’s PART bracis-33583 250 4 overall overall ADJ bracis-33583 250 5 performance performance NOUN bracis-33583 250 6 proved prove VERB bracis-33583 250 7 to to PART bracis-33583 250 8 be be AUX bracis-33583 250 9 an an DET bracis-33583 250 10 improvement improvement NOUN bracis-33583 250 11 over over ADP bracis-33583 250 12 the the DET bracis-33583 250 13 original original ADJ bracis-33583 250 14 som som PROPN bracis-33583 250 15 - - PUNCT bracis-33583 250 16 mll mll PROPN bracis-33583 250 17 . . PUNCT bracis-33583 251 1 the the DET bracis-33583 251 2 high high ADJ bracis-33583 251 3 values value NOUN bracis-33583 251 4 of of ADP bracis-33583 251 5 the the DET bracis-33583 251 6 spread spread ADJ bracis-33583 251 7 factor factor NOUN bracis-33583 251 8 show show VERB bracis-33583 251 9 that that SCONJ bracis-33583 251 10 the the DET bracis-33583 251 11 datasets dataset NOUN bracis-33583 251 12 benefit benefit VERB bracis-33583 251 13 from from ADP bracis-33583 251 14 a a DET bracis-33583 251 15 higher high ADJ bracis-33583 251 16 number number NOUN bracis-33583 251 17 of of ADP bracis-33583 251 18 neurons neuron NOUN bracis-33583 251 19 in in ADP bracis-33583 251 20 a a DET bracis-33583 251 21 dynamic dynamic ADJ bracis-33583 251 22 grid grid NOUN bracis-33583 251 23 . . PUNCT bracis-33583 252 1 unfortunately unfortunately ADV bracis-33583 252 2 , , PUNCT bracis-33583 252 3 gsom gsom PROPN bracis-33583 252 4 - - PUNCT bracis-33583 252 5 mll mll PROPN bracis-33583 252 6 performs perform VERB bracis-33583 252 7 poorly poorly ADV bracis-33583 252 8 in in ADP bracis-33583 252 9 datasets dataset NOUN bracis-33583 252 10 with with ADP bracis-33583 252 11 balanced balanced ADJ bracis-33583 252 12 label label NOUN bracis-33583 252 13 cardinality cardinality NOUN bracis-33583 252 14 and and CCONJ bracis-33583 252 15 small small ADJ bracis-33583 252 16 label label NOUN bracis-33583 252 17 density density NOUN bracis-33583 252 18 with with ADP bracis-33583 252 19 many many ADJ bracis-33583 252 20 distinct distinct ADJ bracis-33583 252 21 label label NOUN bracis-33583 252 22 sets set NOUN bracis-33583 252 23 . . PUNCT bracis-33583 253 1 however however ADV bracis-33583 253 2 , , PUNCT bracis-33583 253 3 it it PRON bracis-33583 253 4 shows show VERB bracis-33583 253 5 consistent consistent ADJ bracis-33583 253 6 and and CCONJ bracis-33583 253 7 competitive competitive ADJ bracis-33583 253 8 results result NOUN bracis-33583 253 9 in in ADP bracis-33583 253 10 other other ADJ bracis-33583 253 11 scenarios scenario NOUN bracis-33583 253 12 . . PUNCT bracis-33583 254 1 in in ADP bracis-33583 254 2 future future ADJ bracis-33583 254 3 works work NOUN bracis-33583 254 4 , , PUNCT bracis-33583 254 5 we we PRON bracis-33583 254 6 can can AUX bracis-33583 254 7 explore explore VERB bracis-33583 254 8 the the DET bracis-33583 254 9 neighborhood neighborhood NOUN bracis-33583 254 10 ’s ’s PART bracis-33583 254 11 influence influence NOUN bracis-33583 254 12 on on ADP bracis-33583 254 13 the the DET bracis-33583 254 14 creation creation NOUN bracis-33583 254 15 of of ADP bracis-33583 254 16 the the DET bracis-33583 254 17 prototype prototype NOUN bracis-33583 254 18 vector vector NOUN bracis-33583 254 19 . . PUNCT bracis-33583 255 1 exploring explore VERB bracis-33583 255 2 different different ADJ bracis-33583 255 3 learning learning NOUN bracis-33583 255 4 rate rate NOUN bracis-33583 255 5 ranges range NOUN bracis-33583 255 6 also also ADV bracis-33583 255 7 allows allow VERB bracis-33583 255 8 for for ADP bracis-33583 255 9 analyzing analyze VERB bracis-33583 255 10 the the DET bracis-33583 255 11 model model NOUN bracis-33583 255 12 ’s ’s PART bracis-33583 255 13 performance performance NOUN bracis-33583 255 14 , , PUNCT bracis-33583 255 15 as as SCONJ bracis-33583 255 16 many many ADJ bracis-33583 255 17 datasets dataset NOUN bracis-33583 255 18 have have VERB bracis-33583 255 19 their their PRON bracis-33583 255 20 optimal optimal ADJ bracis-33583 255 21 value value NOUN bracis-33583 255 22 near near ADP bracis-33583 255 23 the the DET bracis-33583 255 24 limits limit NOUN bracis-33583 255 25 tested test VERB bracis-33583 255 26 when when SCONJ bracis-33583 255 27 considering consider VERB bracis-33583 255 28 error error NOUN bracis-33583 255 29 rates rate NOUN bracis-33583 255 30 . . PUNCT bracis-33583 256 1 another another DET bracis-33583 256 2 way way NOUN bracis-33583 256 3 to to PART bracis-33583 256 4 expand expand VERB bracis-33583 256 5 the the DET bracis-33583 256 6 research research NOUN bracis-33583 256 7 would would AUX bracis-33583 256 8 be be AUX bracis-33583 256 9 to to PART bracis-33583 256 10 utilize utilize VERB bracis-33583 256 11 hierarchical hierarchical ADJ bracis-33583 256 12 growing grow VERB bracis-33583 256 13 self self NOUN bracis-33583 256 14 - - PUNCT bracis-33583 256 15 organizing organize VERB bracis-33583 256 16 maps map NOUN bracis-33583 256 17     SPACE bracis-33583 257 1 [ [ X bracis-33583 257 2 25 25 NUM bracis-33583 257 3 ] ] PUNCT bracis-33583 257 4 . . PUNCT bracis-33583 258 1 this this PRON bracis-33583 258 2 involves involve VERB bracis-33583 258 3 creating create VERB bracis-33583 258 4 a a DET bracis-33583 258 5 hierarchy hierarchy NOUN bracis-33583 258 6 of of ADP bracis-33583 258 7 gsoms gsom NOUN bracis-33583 258 8 based base VERB bracis-33583 258 9 on on ADP bracis-33583 258 10 a a DET bracis-33583 258 11 hierarchical hierarchical ADJ bracis-33583 258 12 threshold threshold NOUN bracis-33583 258 13 as as ADP bracis-33583 258 14 a a DET bracis-33583 258 15 direct direct ADJ bracis-33583 258 16 extension extension NOUN bracis-33583 258 17 of of ADP bracis-33583 258 18 som som NOUN bracis-33583 258 19 and and CCONJ bracis-33583 258 20 gsom gsom PROPN bracis-33583 258 21 . . PUNCT bracis-33583 259 1 notes note NOUN bracis-33583 259 2 1.https://mulan.sourceforge.net/datasets-mlc.html 1.https://mulan.sourceforge.net/datasets-mlc.html NUM bracis-33583 259 3 . . PUNCT bracis-33583 260 1 2.https://github.com/rodrigosales/gsom 2.https://github.com/rodrigosales/gsom NUM bracis-33583 260 2 . . PUNCT bracis-33583 261 1 3.https://github.com/pedrohcasarotto/gsom-mll 3.https://github.com/pedrohcasarotto/gsom-mll NUM bracis-33583 261 2 . . PUNCT bracis-33583 262 1 references reference NOUN bracis-33583 262 2 growingsom growingsom NOUN bracis-33583 262 3 r r NOUN bracis-33583 262 4 package package NOUN bracis-33583 262 5 . . PUNCT bracis-33583 263 1 https://github.com/alexhunziker/growingsom https://github.com/alexhunziker/growingsom NOUN bracis-33583 263 2 gsom gsom VERB bracis-33583 263 3 the the DET bracis-33583 263 4 growing grow VERB bracis-33583 263 5 self self NOUN bracis-33583 263 6 organizing organizing NOUN bracis-33583 263 7 map map NOUN bracis-33583 263 8 implementation implementation NOUN bracis-33583 263 9 on on ADP bracis-33583 263 10 python python PROPN bracis-33583 263 11 . . PUNCT bracis-33583 264 1 https://github.com/anantadata/gsom https://github.com/anantadata/gsom PROPN bracis-33583 264 2 pygsom pygsom PROPN bracis-33583 264 3 a a DET bracis-33583 264 4 gsom gsom NOUN bracis-33583 264 5 ( ( PUNCT bracis-33583 264 6 growing grow VERB bracis-33583 264 7 self self NOUN bracis-33583 264 8 - - PUNCT bracis-33583 264 9 organizing organize VERB bracis-33583 264 10 map map NOUN bracis-33583 264 11 ) ) PUNCT bracis-33583 264 12 implementation implementation NOUN bracis-33583 264 13 for for ADP bracis-33583 264 14 python python PROPN bracis-33583 264 15 . . PUNCT bracis-33583 264 16 https://github.com/philippludwig/pygsom https://github.com/philippludwig/pygsom PROPN bracis-33583 264 17 adeu adeu PROPN bracis-33583 264 18 , , PUNCT bracis-33583 264 19 r.s r.s PROPN bracis-33583 264 20 . . PROPN bracis-33583 264 21 , , PUNCT bracis-33583 264 22 ferreira ferreira PROPN bracis-33583 264 23 , , PUNCT bracis-33583 264 24 k.r k.r PROPN bracis-33583 264 25 . . PROPN bracis-33583 264 26 , , PUNCT bracis-33583 264 27 andrade andrade PROPN bracis-33583 264 28 , , PUNCT bracis-33583 264 29 p.r p.r PROPN bracis-33583 264 30 . . PROPN bracis-33583 264 31 , , PUNCT bracis-33583 264 32 santos santos PROPN bracis-33583 264 33 , , PUNCT bracis-33583 264 34 l. l. PROPN bracis-33583 264 35 : : PUNCT bracis-33583 264 36 evaluating evaluate VERB bracis-33583 264 37 growing grow VERB bracis-33583 264 38 self self NOUN bracis-33583 264 39 - - PUNCT bracis-33583 264 40 organizing organize VERB bracis-33583 264 41 maps map NOUN bracis-33583 264 42 for for ADP bracis-33583 264 43 satellite satellite NOUN bracis-33583 264 44 image image NOUN bracis-33583 264 45 time time NOUN bracis-33583 264 46 series series PROPN bracis-33583 264 47 clustering cluster VERB bracis-33583 264 48 . . PUNCT bracis-33583 265 1 in in ADP bracis-33583 265 2 : : PUNCT bracis-33583 265 3 geoinfo geoinfo NOUN bracis-33583 265 4 , , PUNCT bracis-33583 265 5 20 20 NUM bracis-33583 265 6 years year NOUN bracis-33583 265 7 after after ADV bracis-33583 265 8 ! ! PUNCT bracis-33583 266 1 p. p. NOUN bracis-33583 266 2     SPACE bracis-33583 266 3 243 243 NUM bracis-33583 266 4 ( ( PUNCT bracis-33583 266 5 2019 2019 NUM bracis-33583 266 6 ) ) PUNCT bracis-33583 266 7 google google PROPN bracis-33583 266 8 scholar scholar NOUN bracis-33583 266 9     SPACE bracis-33583 266 10 aha aha PROPN bracis-33583 266 11 , , PUNCT bracis-33583 266 12 d.w d.w PROPN bracis-33583 266 13 . . PROPN bracis-33583 266 14 , , PUNCT bracis-33583 266 15 kibler kibler PROPN bracis-33583 266 16 , , PUNCT bracis-33583 266 17 d. d. PROPN bracis-33583 266 18 , , PUNCT bracis-33583 266 19 albert albert PROPN bracis-33583 266 20 , , PUNCT bracis-33583 266 21 m.k m.k PROPN bracis-33583 266 22 . . PUNCT bracis-33583 266 23 : : PUNCT bracis-33583 267 1 instance instance NOUN bracis-33583 267 2 - - PUNCT bracis-33583 267 3 based base VERB bracis-33583 267 4 learning learning NOUN bracis-33583 267 5 algorithms algorithm NOUN bracis-33583 267 6 . . PUNCT bracis-33583 268 1 mach mach NOUN bracis-33583 268 2 . . PUNCT bracis-33583 269 1 learn learn VERB bracis-33583 269 2 . . PUNCT bracis-33583 270 1 6(1 6(1 NUM bracis-33583 270 2 ) ) PUNCT bracis-33583 271 1 , , PUNCT bracis-33583 271 2 37–66 37–66 NUM bracis-33583 271 3 ( ( PUNCT bracis-33583 271 4 1991 1991 NUM bracis-33583 271 5 ) ) PUNCT bracis-33583 271 6 article article NOUN bracis-33583 271 7     SPACE bracis-33583 271 8 math math NOUN bracis-33583 271 9     SPACE bracis-33583 271 10 google google PROPN bracis-33583 271 11 scholar scholar NOUN bracis-33583 271 12     SPACE bracis-33583 271 13 alahakoon alahakoon NOUN bracis-33583 271 14 , , PUNCT bracis-33583 271 15 d. d. PROPN bracis-33583 271 16 , , PUNCT bracis-33583 271 17 halgamuge halgamuge NOUN bracis-33583 271 18 , , PUNCT bracis-33583 271 19 s.k s.k PROPN bracis-33583 271 20 . . PROPN bracis-33583 271 21 , , PUNCT bracis-33583 271 22 srinivasan srinivasan NOUN bracis-33583 271 23 , , PUNCT bracis-33583 271 24 b. b. NOUN bracis-33583 271 25 : : PUNCT bracis-33583 272 1 dynamic dynamic ADJ bracis-33583 272 2 self self NOUN bracis-33583 272 3 - - PUNCT bracis-33583 272 4 organizing organize VERB bracis-33583 272 5 maps map NOUN bracis-33583 272 6 with with ADP bracis-33583 272 7 controlled control VERB bracis-33583 272 8 growth growth NOUN bracis-33583 272 9 for for ADP bracis-33583 272 10 knowledge knowledge NOUN bracis-33583 272 11 discovery discovery NOUN bracis-33583 272 12 . . PUNCT bracis-33583 273 1 ieee ieee PROPN bracis-33583 273 2 trans trans PROPN bracis-33583 273 3 . . PUNCT bracis-33583 274 1 neural neural ADJ bracis-33583 274 2 netw netw NOUN bracis-33583 274 3 . . PUNCT bracis-33583 275 1 11(3 11(3 NUM bracis-33583 275 2 ) ) PUNCT bracis-33583 275 3 , , PUNCT bracis-33583 275 4 601–614 601–614 NUM bracis-33583 275 5 ( ( PUNCT bracis-33583 275 6 2000 2000 NUM bracis-33583 275 7 ) ) PUNCT bracis-33583 275 8 article article NOUN bracis-33583 275 9     SPACE bracis-33583 275 10 math math NOUN bracis-33583 275 11     SPACE bracis-33583 275 12 google google PROPN bracis-33583 275 13 scholar scholar NOUN bracis-33583 275 14     SPACE bracis-33583 275 15 alshanqiti alshanqiti PROPN bracis-33583 275 16 , , PUNCT bracis-33583 275 17 a. a. NOUN bracis-33583 275 18 , , PUNCT bracis-33583 275 19 namoun namoun PROPN bracis-33583 275 20 , , PUNCT bracis-33583 275 21 a. a. NOUN bracis-33583 275 22 : : PUNCT bracis-33583 275 23 predicting predict VERB bracis-33583 275 24 student student NOUN bracis-33583 275 25 performance performance NOUN bracis-33583 275 26 and and CCONJ bracis-33583 275 27 its its PRON bracis-33583 275 28 influential influential ADJ bracis-33583 275 29 factors factor NOUN bracis-33583 275 30 using use VERB bracis-33583 275 31 hybrid hybrid ADJ bracis-33583 275 32 regression regression NOUN bracis-33583 275 33 and and CCONJ bracis-33583 275 34 multi multi ADJ bracis-33583 275 35 - - ADJ bracis-33583 275 36 label label ADJ bracis-33583 275 37 classification classification NOUN bracis-33583 275 38 . . PUNCT bracis-33583 276 1 ieee ieee NOUN bracis-33583 276 2 access access NOUN bracis-33583 276 3 8 8 NUM bracis-33583 276 4 , , PUNCT bracis-33583 276 5 203827–203844 203827–203844 NUM bracis-33583 276 6 ( ( PUNCT bracis-33583 276 7 2020 2020 NUM bracis-33583 276 8 ) ) PUNCT bracis-33583 276 9 article article NOUN bracis-33583 276 10     SPACE bracis-33583 276 11 google google PROPN bracis-33583 276 12 scholar scholar NOUN bracis-33583 276 13     SPACE bracis-33583 276 14 baltruschat baltruschat NOUN bracis-33583 276 15 , , PUNCT bracis-33583 276 16 i.m i.m PROPN bracis-33583 276 17 . . PROPN bracis-33583 276 18 , , PUNCT bracis-33583 276 19 nickisch nickisch PROPN bracis-33583 276 20 , , PUNCT bracis-33583 276 21 h. h. NOUN bracis-33583 276 22 , , PUNCT bracis-33583 276 23 grass grass NOUN bracis-33583 276 24 , , PUNCT bracis-33583 276 25 m. m. NOUN bracis-33583 276 26 , , PUNCT bracis-33583 276 27 knopp knopp ADJ bracis-33583 276 28 , , PUNCT bracis-33583 276 29 t. t. PROPN bracis-33583 276 30 , , PUNCT bracis-33583 276 31 saalbach saalbach PROPN bracis-33583 276 32 , , PUNCT bracis-33583 276 33 a. a. NOUN bracis-33583 276 34 : : PUNCT bracis-33583 276 35 comparison comparison NOUN bracis-33583 276 36 of of ADP bracis-33583 276 37 deep deep ADJ bracis-33583 276 38 learning learning NOUN bracis-33583 276 39 approaches approach NOUN bracis-33583 276 40 for for ADP bracis-33583 276 41 multi multi ADJ bracis-33583 276 42 - - ADJ bracis-33583 276 43 label label ADJ bracis-33583 276 44 chest chest NOUN bracis-33583 276 45 x x NOUN bracis-33583 276 46 - - NOUN bracis-33583 276 47 ray ray NOUN bracis-33583 276 48 classification classification NOUN bracis-33583 276 49 . . PUNCT bracis-33583 277 1 sci sci PROPN bracis-33583 277 2 . . PROPN bracis-33583 277 3 rep rep PROPN bracis-33583 277 4 . . PROPN bracis-33583 277 5 9(1 9(1 NUM bracis-33583 277 6 ) ) PUNCT bracis-33583 277 7 , , PUNCT bracis-33583 277 8 1–10 1–10 NOUN bracis-33583 277 9 ( ( PUNCT bracis-33583 277 10 2019 2019 NUM bracis-33583 277 11 ) ) PUNCT bracis-33583 277 12 article article NOUN bracis-33583 277 13     SPACE bracis-33583 277 14 google google PROPN bracis-33583 277 15 scholar scholar NOUN bracis-33583 277 16     SPACE bracis-33583 277 17 banerjee banerjee PROPN bracis-33583 277 18 , , PUNCT bracis-33583 277 19 s. s. PROPN bracis-33583 277 20 , , PUNCT bracis-33583 277 21 akkaya akkaya PROPN bracis-33583 277 22 , , PUNCT bracis-33583 277 23 c. c. PROPN bracis-33583 277 24 , , PUNCT bracis-33583 277 25 perez perez PROPN bracis-33583 277 26 - - PUNCT bracis-33583 277 27 sorrosal sorrosal NOUN bracis-33583 277 28 , , PUNCT bracis-33583 277 29 f. f. PROPN bracis-33583 277 30 , , PUNCT bracis-33583 277 31 tsioutsiouliklis tsioutsiouliklis PROPN bracis-33583 277 32 , , PUNCT bracis-33583 277 33 k. k. PROPN bracis-33583 277 34 : : PUNCT bracis-33583 277 35 hierarchical hierarchical ADJ bracis-33583 277 36 transfer transfer NOUN bracis-33583 277 37 learning learning NOUN bracis-33583 277 38 for for ADP bracis-33583 277 39 multi multi ADJ bracis-33583 277 40 - - ADJ bracis-33583 277 41 label label ADJ bracis-33583 277 42 text text NOUN bracis-33583 277 43 classification classification NOUN bracis-33583 277 44 . . PUNCT bracis-33583 278 1 in in ADP bracis-33583 278 2 : : PUNCT bracis-33583 278 3 proceedings proceeding NOUN bracis-33583 278 4 of of ADP bracis-33583 278 5 the the DET bracis-33583 278 6 57th 57th ADJ bracis-33583 278 7 annual annual ADJ bracis-33583 278 8 meeting meeting NOUN bracis-33583 278 9 of of ADP bracis-33583 278 10 the the DET bracis-33583 278 11 association association NOUN bracis-33583 278 12 for for ADP bracis-33583 278 13 computational computational ADJ bracis-33583 278 14 linguistics linguistic NOUN bracis-33583 278 15 , , PUNCT bracis-33583 278 16 pp pp ADJ bracis-33583 278 17 . . PUNCT bracis-33583 279 1 6295–6300 6295–6300 NUM bracis-33583 279 2 ( ( PUNCT bracis-33583 279 3 2019 2019 NUM bracis-33583 279 4 ) ) PUNCT bracis-33583 279 5 google google NOUN bracis-33583 279 6 scholar scholar NOUN bracis-33583 279 7     SPACE bracis-33583 279 8 blockeel blockeel NOUN bracis-33583 279 9 , , PUNCT bracis-33583 279 10 h. h. PROPN bracis-33583 279 11 , , PUNCT bracis-33583 279 12 raedt raedt NOUN bracis-33583 279 13 , , PUNCT bracis-33583 279 14 l.d l.d PROPN bracis-33583 279 15 . . PROPN bracis-33583 279 16 , , PUNCT bracis-33583 279 17 ramon ramon PROPN bracis-33583 279 18 , , PUNCT bracis-33583 279 19 j. j. PROPN bracis-33583 279 20 : : PUNCT bracis-33583 279 21 top top ADJ bracis-33583 279 22 - - PUNCT bracis-33583 279 23 down down ADP bracis-33583 279 24 induction induction NOUN bracis-33583 279 25 of of ADP bracis-33583 279 26 clustering cluster VERB bracis-33583 279 27 trees tree NOUN bracis-33583 279 28 . . PUNCT bracis-33583 280 1 in in ADP bracis-33583 280 2 : : PUNCT bracis-33583 280 3 proceedings proceeding NOUN bracis-33583 280 4 of of ADP bracis-33583 280 5 the the DET bracis-33583 280 6 fifteenth fifteenth ADJ bracis-33583 280 7 international international ADJ bracis-33583 280 8 conference conference NOUN bracis-33583 280 9 on on ADP bracis-33583 280 10 machine machine NOUN bracis-33583 280 11 learning learning NOUN bracis-33583 280 12 , , PUNCT bracis-33583 280 13 icml icml VERB bracis-33583 280 14 1998 1998 NUM bracis-33583 280 15 , , PUNCT bracis-33583 280 16 pp pp ADP bracis-33583 280 17 . . PUNCT bracis-33583 281 1 55–63 55–63 NUM bracis-33583 281 2 . . PUNCT bracis-33583 282 1 morgan morgan PROPN bracis-33583 282 2 kaufmann kaufmann PROPN bracis-33583 282 3 publishers publishers PROPN bracis-33583 282 4 inc inc PROPN bracis-33583 282 5 . . PROPN bracis-33583 282 6 , , PUNCT bracis-33583 282 7 san san PROPN bracis-33583 282 8 francisco francisco PROPN bracis-33583 282 9 , , PUNCT bracis-33583 282 10 ca ca PROPN bracis-33583 282 11 , , PUNCT bracis-33583 282 12 usa usa PROPN bracis-33583 282 13 ( ( PUNCT bracis-33583 282 14 1998 1998 NUM bracis-33583 282 15 ) ) PUNCT bracis-33583 282 16 google google PROPN bracis-33583 282 17 scholar scholar NOUN bracis-33583 282 18     SPACE bracis-33583 282 19 boutell boutell NOUN bracis-33583 282 20 , , PUNCT bracis-33583 282 21 m.r m.r PROPN bracis-33583 282 22 . . PROPN bracis-33583 282 23 , , PUNCT bracis-33583 282 24 luo luo PROPN bracis-33583 282 25 , , PUNCT bracis-33583 282 26 j. j. PROPN bracis-33583 282 27 , , PUNCT bracis-33583 282 28 shen shen PROPN bracis-33583 282 29 , , PUNCT bracis-33583 282 30 x. x. PROPN bracis-33583 282 31 , , PUNCT bracis-33583 282 32 brown brown PROPN bracis-33583 282 33 , , PUNCT bracis-33583 282 34 c.m c.m PROPN bracis-33583 282 35 . . PUNCT bracis-33583 282 36 : : PUNCT bracis-33583 282 37 learning learn VERB bracis-33583 282 38 multi multi ADJ bracis-33583 282 39 - - ADJ bracis-33583 282 40 label label ADJ bracis-33583 282 41 scene scene NOUN bracis-33583 282 42 classification classification NOUN bracis-33583 282 43 . . PUNCT bracis-33583 283 1 pattern pattern NOUN bracis-33583 283 2 recogn recogn VERB bracis-33583 283 3 . . PUNCT bracis-33583 284 1 37(9 37(9 NUM bracis-33583 284 2 ) ) PUNCT bracis-33583 284 3 , , PUNCT bracis-33583 284 4 1757–1771 1757–1771 NUM bracis-33583 284 5 ( ( PUNCT bracis-33583 284 6 2004 2004 NUM bracis-33583 284 7 ) ) PUNCT bracis-33583 284 8 article article NOUN bracis-33583 284 9     SPACE bracis-33583 284 10 math math NOUN bracis-33583 284 11     SPACE bracis-33583 284 12 google google PROPN bracis-33583 284 13 scholar scholar NOUN bracis-33583 284 14     SPACE bracis-33583 284 15 chen chen PROPN bracis-33583 284 16 , , PUNCT bracis-33583 284 17 q. q. PROPN bracis-33583 284 18 , , PUNCT bracis-33583 284 19 et et PROPN bracis-33583 284 20     SPACE bracis-33583 284 21 al al PROPN bracis-33583 284 22 . . PROPN bracis-33583 284 23 : : PUNCT bracis-33583 284 24 multi multi ADJ bracis-33583 284 25 - - ADJ bracis-33583 284 26 label label ADJ bracis-33583 284 27 classification classification NOUN bracis-33583 284 28 for for ADP bracis-33583 284 29 biomedical biomedical ADJ bracis-33583 284 30 literature literature NOUN bracis-33583 284 31 : : PUNCT bracis-33583 284 32 an an DET bracis-33583 284 33 overview overview NOUN bracis-33583 284 34 of of ADP bracis-33583 284 35 the the DET bracis-33583 284 36 biocreative biocreative ADJ bracis-33583 284 37 vii vii PROPN bracis-33583 284 38 litcovid litcovid PROPN bracis-33583 284 39 track track NOUN bracis-33583 284 40 for for ADP bracis-33583 284 41 covid-19 covid-19 PROPN bracis-33583 284 42 literature literature NOUN bracis-33583 284 43 topic topic NOUN bracis-33583 284 44 annotations annotation NOUN bracis-33583 284 45 . . PUNCT bracis-33583 285 1 database database NOUN bracis-33583 285 2 2022 2022 NUM bracis-33583 285 3 , , PUNCT bracis-33583 285 4 baac069 baac069 PROPN bracis-33583 285 5 ( ( PUNCT bracis-33583 285 6 2022 2022 NUM bracis-33583 285 7 ) ) PUNCT bracis-33583 285 8 google google PROPN bracis-33583 285 9 scholar scholar NOUN bracis-33583 285 10     SPACE bracis-33583 285 11 cheng cheng PROPN bracis-33583 285 12 , , PUNCT bracis-33583 285 13 w. w. PROPN bracis-33583 285 14 , , PUNCT bracis-33583 285 15 hüllermeier hüllermeier PROPN bracis-33583 285 16 , , PUNCT bracis-33583 285 17 e. e. PROPN bracis-33583 285 18 : : PUNCT bracis-33583 285 19 combining combine VERB bracis-33583 285 20 instance instance NOUN bracis-33583 285 21 - - PUNCT bracis-33583 285 22 based base VERB bracis-33583 285 23 learning learning NOUN bracis-33583 285 24 and and CCONJ bracis-33583 285 25 logistic logistic ADJ bracis-33583 285 26 regression regression NOUN bracis-33583 285 27 for for ADP bracis-33583 285 28 multilabel multilabel NOUN bracis-33583 285 29 classification classification NOUN bracis-33583 285 30 . . PUNCT bracis-33583 286 1 mach mach NOUN bracis-33583 286 2 . . PUNCT bracis-33583 287 1 learn learn VERB bracis-33583 287 2 . . PUNCT bracis-33583 288 1 76(2–3 76(2–3 NUM bracis-33583 288 2 ) ) PUNCT bracis-33583 288 3 , , PUNCT bracis-33583 288 4 211–225 211–225 NUM bracis-33583 288 5 ( ( PUNCT bracis-33583 288 6 2009 2009 NUM bracis-33583 288 7 ) ) PUNCT bracis-33583 288 8 article article NOUN bracis-33583 288 9     SPACE bracis-33583 288 10 math math NOUN bracis-33583 288 11     SPACE bracis-33583 288 12 google google PROPN bracis-33583 288 13 scholar scholar NOUN bracis-33583 288 14     SPACE bracis-33583 288 15 cheng cheng PROPN bracis-33583 288 16 , , PUNCT bracis-33583 288 17 w. w. PROPN bracis-33583 288 18 , , PUNCT bracis-33583 288 19 hüllermeier hüllermeier PROPN bracis-33583 288 20 , , PUNCT bracis-33583 288 21 e. e. PROPN bracis-33583 288 22 , , PUNCT bracis-33583 288 23 dembczynski dembczynski PROPN bracis-33583 288 24 , , PUNCT bracis-33583 288 25 k.j k.j PROPN bracis-33583 288 26 . . PUNCT bracis-33583 288 27 : : PUNCT bracis-33583 289 1 bayes bayes PROPN bracis-33583 289 2 optimal optimal ADJ bracis-33583 289 3 multilabel multilabel NOUN bracis-33583 289 4 classification classification NOUN bracis-33583 289 5 via via ADP bracis-33583 289 6 probabilistic probabilistic ADJ bracis-33583 289 7 classifier classifier NOUN bracis-33583 289 8 chains chain NOUN bracis-33583 289 9 . . PUNCT bracis-33583 290 1 in in ADP bracis-33583 290 2 : : PUNCT bracis-33583 290 3 proceedings proceeding NOUN bracis-33583 290 4 of of ADP bracis-33583 290 5 the the DET bracis-33583 290 6 27th 27th ADJ bracis-33583 290 7 international international ADJ bracis-33583 290 8 conference conference NOUN bracis-33583 290 9 on on ADP bracis-33583 290 10 machine machine NOUN bracis-33583 290 11 learning learning NOUN bracis-33583 290 12 , , PUNCT bracis-33583 290 13 icml-10 icml-10 PROPN bracis-33583 290 14 , , PUNCT bracis-33583 290 15 pp pp ADJ bracis-33583 290 16 . . PUNCT bracis-33583 291 1 279–286 279–286 NUM bracis-33583 291 2 ( ( PUNCT bracis-33583 291 3 2010 2010 NUM bracis-33583 291 4 ) ) PUNCT bracis-33583 291 5 google google PROPN bracis-33583 291 6 scholar scholar NOUN bracis-33583 291 7     SPACE bracis-33583 291 8 cherman cherman PROPN bracis-33583 291 9 , , PUNCT bracis-33583 291 10 e.a e.a PROPN bracis-33583 291 11 . . PROPN bracis-33583 291 12 : : PUNCT bracis-33583 292 1 aprendizado aprendizado PROPN bracis-33583 292 2 de de PROPN bracis-33583 292 3 máquina máquina PROPN bracis-33583 292 4 multirrótulo multirrótulo PROPN bracis-33583 292 5 : : PUNCT bracis-33583 292 6 explorando explorando PROPN bracis-33583 292 7 a a DET bracis-33583 292 8 dependência dependência PROPN bracis-33583 292 9 de de ADP bracis-33583 292 10 rótulos rótulos PROPN bracis-33583 292 11 e e X bracis-33583 292 12 o o PROPN bracis-33583 292 13 aprendizado aprendizado PROPN bracis-33583 292 14 ativo ativo PROPN bracis-33583 292 15 . . PUNCT bracis-33583 293 1 ph.d ph.d PROPN bracis-33583 293 2 . . PUNCT bracis-33583 294 1 thesis thesis PROPN bracis-33583 294 2 , , PUNCT bracis-33583 294 3 universidade universidade PROPN bracis-33583 294 4 de de PROPN bracis-33583 294 5 são são PROPN bracis-33583 294 6 paulo paulo PROPN bracis-33583 294 7 ( ( PUNCT bracis-33583 294 8 2013 2013 NUM bracis-33583 294 9 ) ) PUNCT bracis-33583 294 10 google google PROPN bracis-33583 294 11 scholar scholar NOUN bracis-33583 294 12     SPACE bracis-33583 294 13 clare clare NOUN bracis-33583 294 14 , , PUNCT bracis-33583 294 15 a. a. NOUN bracis-33583 294 16 , , PUNCT bracis-33583 294 17 king king NOUN bracis-33583 294 18 , , PUNCT bracis-33583 294 19 r.d r.d PROPN bracis-33583 294 20 . . PUNCT bracis-33583 294 21 : : PUNCT bracis-33583 295 1 knowledge knowledge NOUN bracis-33583 295 2 discovery discovery PROPN bracis-33583 295 3 in in ADP bracis-33583 295 4 multi multi ADJ bracis-33583 295 5 - - ADJ bracis-33583 295 6 label label ADJ bracis-33583 295 7 phenotype phenotype NOUN bracis-33583 295 8 data datum NOUN bracis-33583 295 9 . . PUNCT bracis-33583 296 1 in in ADP bracis-33583 296 2 : : PUNCT bracis-33583 296 3 de de X bracis-33583 296 4 raedt raedt NOUN bracis-33583 296 5 , , PUNCT bracis-33583 296 6 l. l. PROPN bracis-33583 296 7 , , PUNCT bracis-33583 296 8 siebes siebe NOUN bracis-33583 296 9 , , PUNCT bracis-33583 296 10 a. a. NOUN bracis-33583 296 11 ( ( PUNCT bracis-33583 296 12 eds eds PROPN bracis-33583 296 13 . . PUNCT bracis-33583 296 14 ) ) PUNCT bracis-33583 296 15 pkdd pkdd VERB bracis-33583 296 16 2001 2001 NUM bracis-33583 296 17 . . PUNCT bracis-33583 297 1 lncs lncs PROPN bracis-33583 297 2 ( ( PUNCT bracis-33583 297 3 lnai lnai ADJ bracis-33583 297 4 ) ) PUNCT bracis-33583 297 5 , , PUNCT bracis-33583 297 6 vol vol NOUN bracis-33583 297 7 . . PUNCT bracis-33583 297 8 2168 2168 NUM bracis-33583 297 9 , , PUNCT bracis-33583 297 10 pp pp ADJ bracis-33583 297 11 . . PUNCT bracis-33583 298 1 42–53 42–53 NUM bracis-33583 298 2 . . PUNCT bracis-33583 298 3 springer springer NOUN bracis-33583 298 4 , , PUNCT bracis-33583 298 5 heidelberg heidelberg PROPN bracis-33583 298 6 ( ( PUNCT bracis-33583 298 7 2001 2001 NUM bracis-33583 298 8 ) ) PUNCT bracis-33583 298 9 . . PUNCT bracis-33583 299 1 https://doi.org/10.1007/3-540-44794-6_4 https://doi.org/10.1007/3-540-44794-6_4 PRON bracis-33583 299 2 chapter chapter NOUN bracis-33583 299 3     SPACE bracis-33583 299 4 math math NOUN bracis-33583 299 5     SPACE bracis-33583 299 6 google google PROPN bracis-33583 299 7 scholar scholar NOUN bracis-33583 299 8     SPACE bracis-33583 299 9 colombini colombini PROPN bracis-33583 299 10 , , PUNCT bracis-33583 299 11 g.g g.g PROPN bracis-33583 299 12 . . PROPN bracis-33583 299 13 , , PUNCT bracis-33583 299 14 de de NOUN bracis-33583 299 15     SPACE bracis-33583 299 16 abreu abreu NOUN bracis-33583 299 17 , , PUNCT bracis-33583 299 18 i.b.m i.b.m ADJ bracis-33583 299 19 . . PROPN bracis-33583 299 20 , , PUNCT bracis-33583 299 21 cerri cerri PROPN bracis-33583 299 22 , , PUNCT bracis-33583 299 23 r. r. PROPN bracis-33583 299 24 : : PUNCT bracis-33583 299 25 a a DET bracis-33583 299 26 self self NOUN bracis-33583 299 27 - - PUNCT bracis-33583 299 28 organizing organize VERB bracis-33583 299 29 map map NOUN bracis-33583 299 30 - - PUNCT bracis-33583 299 31 based base VERB bracis-33583 299 32 method method NOUN bracis-33583 299 33 for for ADP bracis-33583 299 34 multi multi ADJ bracis-33583 299 35 - - ADJ bracis-33583 299 36 label label ADJ bracis-33583 299 37 classification classification NOUN bracis-33583 299 38 . . PUNCT bracis-33583 300 1 in in ADP bracis-33583 300 2 : : PUNCT bracis-33583 300 3 2017 2017 NUM bracis-33583 300 4 international international ADJ bracis-33583 300 5 joint joint ADJ bracis-33583 300 6 conference conference NOUN bracis-33583 300 7 on on ADP bracis-33583 300 8 neural neural ADJ bracis-33583 300 9 networks network NOUN bracis-33583 300 10 ( ( PUNCT bracis-33583 300 11 ijcnn ijcnn PROPN bracis-33583 300 12 ) ) PUNCT bracis-33583 300 13 , , PUNCT bracis-33583 300 14 pp pp PROPN bracis-33583 300 15 . . PUNCT bracis-33583 300 16 4291–4298 4291–4298 NUM bracis-33583 300 17 . . PUNCT bracis-33583 301 1 ieee ieee NOUN bracis-33583 301 2 ( ( PUNCT bracis-33583 301 3 2017 2017 NUM bracis-33583 301 4 ) ) PUNCT bracis-33583 301 5 google google PROPN bracis-33583 301 6 scholar scholar NOUN bracis-33583 301 7     SPACE bracis-33583 301 8 davis davis PROPN bracis-33583 301 9 , , PUNCT bracis-33583 301 10 j. j. PROPN bracis-33583 301 11 , , PUNCT bracis-33583 301 12 goadrich goadrich NOUN bracis-33583 301 13 , , PUNCT bracis-33583 301 14 m. m. NOUN bracis-33583 301 15 : : PUNCT bracis-33583 301 16 the the DET bracis-33583 301 17 relationship relationship NOUN bracis-33583 301 18 between between ADP bracis-33583 301 19 precision precision NOUN bracis-33583 301 20 - - PUNCT bracis-33583 301 21 recall recall NOUN bracis-33583 301 22 and and CCONJ bracis-33583 301 23 roc roc PROPN bracis-33583 301 24 curves curve NOUN bracis-33583 301 25 . . PUNCT bracis-33583 302 1 in in ADP bracis-33583 302 2 : : PUNCT bracis-33583 302 3 international international ADJ bracis-33583 302 4 conference conference NOUN bracis-33583 302 5 on on ADP bracis-33583 302 6 machine machine NOUN bracis-33583 302 7 learning learning NOUN bracis-33583 302 8 , , PUNCT bracis-33583 302 9 pp pp ADP bracis-33583 302 10 . . PUNCT bracis-33583 303 1 233–240 233–240 NUM bracis-33583 303 2 ( ( PUNCT bracis-33583 303 3 2006 2006 NUM bracis-33583 303 4 ) ) PUNCT bracis-33583 303 5 google google PROPN bracis-33583 303 6 scholar scholar NOUN bracis-33583 303 7     SPACE bracis-33583 303 8 gatto gatto NOUN bracis-33583 303 9 , , PUNCT bracis-33583 303 10 e.c e.c PROPN bracis-33583 303 11 . . PROPN bracis-33583 303 12 , , PUNCT bracis-33583 303 13 valejo valejo ADJ bracis-33583 303 14 , , PUNCT bracis-33583 303 15 a.d.b a.d.b NOUN bracis-33583 303 16 . . NOUN bracis-33583 303 17 , , PUNCT bracis-33583 303 18 ferrandin ferrandin PROPN bracis-33583 303 19 , , PUNCT bracis-33583 303 20 m. m. NOUN bracis-33583 303 21 , , PUNCT bracis-33583 303 22 cerri cerri PROPN bracis-33583 303 23 , , PUNCT bracis-33583 303 24 r. r. PROPN bracis-33583 303 25 : : PUNCT bracis-33583 303 26 community community NOUN bracis-33583 303 27 detection detection NOUN bracis-33583 303 28 for for ADP bracis-33583 303 29 multi multi ADJ bracis-33583 303 30 - - ADJ bracis-33583 303 31 label label ADJ bracis-33583 303 32 classification classification NOUN bracis-33583 303 33 . . PUNCT bracis-33583 304 1 in in ADP bracis-33583 304 2 : : PUNCT bracis-33583 304 3 naldi naldi PROPN bracis-33583 304 4 , , PUNCT bracis-33583 304 5 m.c m.c PROPN bracis-33583 304 6 . . PROPN bracis-33583 304 7 , , PUNCT bracis-33583 304 8 bianchi bianchi PROPN bracis-33583 304 9 , , PUNCT bracis-33583 304 10 r.a.c r.a.c ADP bracis-33583 304 11 . . PUNCT bracis-33583 305 1 ( ( PUNCT bracis-33583 305 2 eds ed NOUN bracis-33583 305 3 . . PUNCT bracis-33583 305 4 ) ) PUNCT bracis-33583 306 1 intelligent intelligent ADJ bracis-33583 306 2 systems system NOUN bracis-33583 306 3 , , PUNCT bracis-33583 306 4 bracis bracis NOUN bracis-33583 306 5 2023 2023 NUM bracis-33583 306 6 . . PUNCT bracis-33583 307 1 lncs lncs PROPN bracis-33583 307 2 , , PUNCT bracis-33583 307 3 vol vol NOUN bracis-33583 307 4 . . PROPN bracis-33583 307 5 14195 14195 NUM bracis-33583 307 6 , , PUNCT bracis-33583 307 7 pp pp ADV bracis-33583 307 8 . . PUNCT bracis-33583 308 1 78–93 78–93 NUM bracis-33583 308 2 . . PUNCT bracis-33583 309 1 springer springer NOUN bracis-33583 309 2 , , PUNCT bracis-33583 309 3 cham cham PROPN bracis-33583 309 4 . . PUNCT bracis-33583 310 1 https://doi.org/10.1007/978-3-031-45368-7_6 https://doi.org/10.1007/978-3-031-45368-7_6 PRON bracis-33583 310 2 godbole godbole NOUN bracis-33583 310 3 , , PUNCT bracis-33583 310 4 s. s. PROPN bracis-33583 310 5 , , PUNCT bracis-33583 310 6 sarawagi sarawagi PROPN bracis-33583 310 7 , , PUNCT bracis-33583 310 8 s. s. PROPN bracis-33583 310 9 : : PUNCT bracis-33583 310 10 discriminative discriminative NOUN bracis-33583 310 11 methods method NOUN bracis-33583 310 12 for for ADP bracis-33583 310 13 multi multi ADJ bracis-33583 310 14 - - ADJ bracis-33583 310 15 labeled label VERB bracis-33583 310 16 classification classification NOUN bracis-33583 310 17 . . PUNCT bracis-33583 311 1 in in ADP bracis-33583 311 2 : : PUNCT bracis-33583 311 3 dai dai PROPN bracis-33583 311 4 , , PUNCT bracis-33583 311 5 h. h. PROPN bracis-33583 311 6 , , PUNCT bracis-33583 311 7 srikant srikant PROPN bracis-33583 311 8 , , PUNCT bracis-33583 311 9 r. r. PROPN bracis-33583 311 10 , , PUNCT bracis-33583 311 11 zhang zhang PROPN bracis-33583 311 12 , , PUNCT bracis-33583 311 13 c. c. PROPN bracis-33583 311 14 ( ( PUNCT bracis-33583 311 15 eds eds PROPN bracis-33583 311 16 . . PUNCT bracis-33583 311 17 ) ) PUNCT bracis-33583 311 18 pakdd pakdd VERB bracis-33583 311 19 2004 2004 NUM bracis-33583 311 20 . . PUNCT bracis-33583 312 1 lncs lncs PROPN bracis-33583 312 2 ( ( PUNCT bracis-33583 312 3 lnai lnai ADJ bracis-33583 312 4 ) ) PUNCT bracis-33583 312 5 , , PUNCT bracis-33583 312 6 vol vol NOUN bracis-33583 312 7 . . PROPN bracis-33583 312 8 3056 3056 NUM bracis-33583 312 9 , , PUNCT bracis-33583 312 10 pp pp ADJ bracis-33583 312 11 . . PUNCT bracis-33583 313 1 22–30 22–30 NUM bracis-33583 313 2 . . PUNCT bracis-33583 314 1 springer springer PROPN bracis-33583 314 2 , , PUNCT bracis-33583 314 3 heidelberg heidelberg PROPN bracis-33583 314 4 ( ( PUNCT bracis-33583 314 5 2004 2004 NUM bracis-33583 314 6 ) ) PUNCT bracis-33583 314 7 . . PUNCT bracis-33583 315 1 https://doi.org/10.1007/978-3-540-24775-3_5 https://doi.org/10.1007/978-3-540-24775-3_5 PROPN bracis-33583 315 2 chapter chapter NOUN bracis-33583 315 3     SPACE bracis-33583 315 4 math math NOUN bracis-33583 315 5     SPACE bracis-33583 315 6 google google PROPN bracis-33583 315 7 scholar scholar NOUN bracis-33583 315 8     SPACE bracis-33583 315 9 kohonen kohonen PROPN bracis-33583 315 10 , , PUNCT bracis-33583 315 11 t.k t.k PROPN bracis-33583 315 12 . . PROPN bracis-33583 315 13 : : PUNCT bracis-33583 316 1 the the DET bracis-33583 316 2 self self NOUN bracis-33583 316 3 - - PUNCT bracis-33583 316 4 organizing organize VERB bracis-33583 316 5 map map NOUN bracis-33583 316 6 . . PUNCT bracis-33583 317 1 proc proc PROPN bracis-33583 317 2 . . PUNCT bracis-33583 318 1 ieee ieee PROPN bracis-33583 318 2 78(9 78(9 NOUN bracis-33583 318 3 ) ) PUNCT bracis-33583 318 4 , , PUNCT bracis-33583 318 5 1464–1480 1464–1480 NUM bracis-33583 318 6 ( ( PUNCT bracis-33583 318 7 1990 1990 NUM bracis-33583 318 8 ) ) PUNCT bracis-33583 318 9 article article NOUN bracis-33583 318 10     SPACE bracis-33583 318 11 math math NOUN bracis-33583 318 12     SPACE bracis-33583 318 13 google google PROPN bracis-33583 318 14 scholar scholar NOUN bracis-33583 318 15     SPACE bracis-33583 318 16 li li PROPN bracis-33583 318 17 , , PUNCT bracis-33583 318 18 q. q. PROPN bracis-33583 318 19 , , PUNCT bracis-33583 318 20 peng peng PROPN bracis-33583 318 21 , , PUNCT bracis-33583 318 22 x. x. PROPN bracis-33583 318 23 , , PUNCT bracis-33583 318 24 qiao qiao PROPN bracis-33583 318 25 , , PUNCT bracis-33583 318 26 y. y. PROPN bracis-33583 318 27 , , PUNCT bracis-33583 318 28 peng peng PROPN bracis-33583 318 29 , , PUNCT bracis-33583 318 30 q. q. PROPN bracis-33583 318 31 : : PUNCT bracis-33583 318 32 learning learn VERB bracis-33583 318 33 category category NOUN bracis-33583 318 34 correlations correlation NOUN bracis-33583 318 35 for for ADP bracis-33583 318 36 multi multi ADJ bracis-33583 318 37 - - ADJ bracis-33583 318 38 label label ADJ bracis-33583 318 39 image image NOUN bracis-33583 318 40 recognition recognition NOUN bracis-33583 318 41 with with ADP bracis-33583 318 42 graph graph NOUN bracis-33583 318 43 networks network NOUN bracis-33583 318 44 . . PUNCT bracis-33583 319 1 arxiv arxiv PROPN bracis-33583 319 2 preprint preprint NOUN bracis-33583 319 3 arxiv:1909.13005 arxiv:1909.13005 NUM bracis-33583 319 4 ( ( PUNCT bracis-33583 319 5 2019 2019 NUM bracis-33583 319 6 ) ) PUNCT bracis-33583 319 7 papanikolaou papanikolaou NOUN bracis-33583 319 8 , , PUNCT bracis-33583 319 9 y. y. PROPN bracis-33583 319 10 , , PUNCT bracis-33583 319 11 tsoumakas tsoumakas PROPN bracis-33583 319 12 , , PUNCT bracis-33583 319 13 g. g. PROPN bracis-33583 319 14 , , PUNCT bracis-33583 319 15 katakis katakis PROPN bracis-33583 319 16 , , PUNCT bracis-33583 319 17 i. i. NOUN bracis-33583 319 18 : : PUNCT bracis-33583 319 19 hierarchical hierarchical ADJ bracis-33583 319 20 partitioning partitioning NOUN bracis-33583 319 21 of of ADP bracis-33583 319 22 the the DET bracis-33583 319 23 output output NOUN bracis-33583 319 24 space space NOUN bracis-33583 319 25 in in ADP bracis-33583 319 26 multi multi ADJ bracis-33583 319 27 - - ADJ bracis-33583 319 28 label label ADJ bracis-33583 319 29 data datum NOUN bracis-33583 319 30 . . PUNCT bracis-33583 320 1 data datum NOUN bracis-33583 320 2 knowl knowl PROPN bracis-33583 320 3 . . PUNCT bracis-33583 321 1 eng eng PROPN bracis-33583 321 2 . . PROPN bracis-33583 322 1 116 116 NUM bracis-33583 322 2 , , PUNCT bracis-33583 322 3 42–60 42–60 NUM bracis-33583 322 4 ( ( PUNCT bracis-33583 322 5 2018 2018 NUM bracis-33583 322 6 ) ) PUNCT bracis-33583 322 7 article article NOUN bracis-33583 322 8     SPACE bracis-33583 322 9 math math NOUN bracis-33583 322 10     SPACE bracis-33583 322 11 google google PROPN bracis-33583 322 12 scholar scholar NOUN bracis-33583 322 13     SPACE bracis-33583 322 14 quinlan quinlan PROPN bracis-33583 322 15 , , PUNCT bracis-33583 322 16 j.r j.r PROPN bracis-33583 322 17 . . PROPN bracis-33583 322 18 : : PUNCT bracis-33583 323 1 c4.5 c4.5 PROPN bracis-33583 323 2 : : PUNCT bracis-33583 323 3 programs program NOUN bracis-33583 323 4 for for ADP bracis-33583 323 5 machine machine NOUN bracis-33583 323 6 learning learning NOUN bracis-33583 323 7 . . PUNCT bracis-33583 324 1 morgan morgan PROPN bracis-33583 324 2 kaufmann kaufmann PROPN bracis-33583 324 3 publishers publishers PROPN bracis-33583 324 4 inc inc PROPN bracis-33583 324 5 . . PROPN bracis-33583 324 6 , , PUNCT bracis-33583 324 7 san san PROPN bracis-33583 324 8 francisco francisco PROPN bracis-33583 324 9 , , PUNCT bracis-33583 324 10 ca ca PROPN bracis-33583 324 11 , , PUNCT bracis-33583 324 12 usa usa PROPN bracis-33583 324 13 ( ( PUNCT bracis-33583 324 14 1993 1993 NUM bracis-33583 324 15 ) ) PUNCT bracis-33583 324 16 google google NOUN bracis-33583 324 17 scholar scholar NOUN bracis-33583 324 18     SPACE bracis-33583 324 19 rauber rauber PROPN bracis-33583 324 20 , , PUNCT bracis-33583 324 21 a. a. NOUN bracis-33583 324 22 , , PUNCT bracis-33583 324 23 merkl merkl NOUN bracis-33583 324 24 , , PUNCT bracis-33583 324 25 d. d. PROPN bracis-33583 324 26 , , PUNCT bracis-33583 324 27 dittenbach dittenbach ADV bracis-33583 324 28 , , PUNCT bracis-33583 324 29 m. m. NOUN bracis-33583 324 30 : : PUNCT bracis-33583 324 31 the the DET bracis-33583 324 32 growing grow VERB bracis-33583 324 33 hierarchical hierarchical ADJ bracis-33583 324 34 self self NOUN bracis-33583 324 35 - - PUNCT bracis-33583 324 36 organizing organize VERB bracis-33583 324 37 map map NOUN bracis-33583 324 38 : : PUNCT bracis-33583 324 39 exploratory exploratory ADJ bracis-33583 324 40 analysis analysis NOUN bracis-33583 324 41 of of ADP bracis-33583 324 42 high high ADJ bracis-33583 324 43 - - PUNCT bracis-33583 324 44 dimensional dimensional ADJ bracis-33583 324 45 data datum NOUN bracis-33583 324 46 . . PUNCT bracis-33583 325 1 ieee ieee PROPN bracis-33583 325 2 trans trans PROPN bracis-33583 325 3 . . PUNCT bracis-33583 326 1 neural neural ADJ bracis-33583 326 2 netw netw NOUN bracis-33583 326 3 . . PUNCT bracis-33583 327 1 13(6 13(6 NUM bracis-33583 327 2 ) ) PUNCT bracis-33583 327 3 , , PUNCT bracis-33583 327 4 1331–1341 1331–1341 NUM bracis-33583 327 5 ( ( PUNCT bracis-33583 327 6 2002 2002 NUM bracis-33583 327 7 ) ) PUNCT bracis-33583 327 8 article article NOUN bracis-33583 327 9     SPACE bracis-33583 327 10 math math NOUN bracis-33583 327 11     SPACE bracis-33583 327 12 google google PROPN bracis-33583 327 13 scholar scholar NOUN bracis-33583 327 14     SPACE bracis-33583 327 15 read read NOUN bracis-33583 327 16 , , PUNCT bracis-33583 327 17 j. j. PROPN bracis-33583 327 18 , , PUNCT bracis-33583 327 19 pfahringer pfahringer NOUN bracis-33583 327 20 , , PUNCT bracis-33583 327 21 b. b. PROPN bracis-33583 327 22 , , PUNCT bracis-33583 327 23 holmes holmes PROPN bracis-33583 327 24 , , PUNCT bracis-33583 327 25 g. g. PROPN bracis-33583 327 26 : : PUNCT bracis-33583 327 27 multi multi ADJ bracis-33583 327 28 - - ADJ bracis-33583 327 29 label label ADJ bracis-33583 327 30 classification classification NOUN bracis-33583 327 31 using use VERB bracis-33583 327 32 ensembles ensemble NOUN bracis-33583 327 33 of of ADP bracis-33583 327 34 pruned prune VERB bracis-33583 327 35 sets set NOUN bracis-33583 327 36 . . PUNCT bracis-33583 328 1 in in ADP bracis-33583 328 2 : : PUNCT bracis-33583 328 3 2008 2008 NUM bracis-33583 328 4 8th 8th ADJ bracis-33583 328 5 ieee ieee NOUN bracis-33583 328 6 international international ADJ bracis-33583 328 7 conference conference NOUN bracis-33583 328 8 on on ADP bracis-33583 328 9 data datum NOUN bracis-33583 328 10 mining mining NOUN bracis-33583 328 11 , , PUNCT bracis-33583 328 12 pp pp ADJ bracis-33583 328 13 . . PUNCT bracis-33583 328 14 995–1000 995–1000 NUM bracis-33583 328 15 . . PUNCT bracis-33583 328 16 ieee ieee NOUN bracis-33583 328 17 ( ( PUNCT bracis-33583 328 18 2008 2008 NUM bracis-33583 328 19 ) ) PUNCT bracis-33583 328 20 google google PROPN bracis-33583 328 21 scholar scholar NOUN bracis-33583 328 22     SPACE bracis-33583 328 23 read read NOUN bracis-33583 328 24 , , PUNCT bracis-33583 328 25 j. j. PROPN bracis-33583 328 26 , , PUNCT bracis-33583 328 27 pfahringer pfahringer NOUN bracis-33583 328 28 , , PUNCT bracis-33583 328 29 b. b. PROPN bracis-33583 328 30 , , PUNCT bracis-33583 328 31 holmes holmes PROPN bracis-33583 328 32 , , PUNCT bracis-33583 328 33 g. g. PROPN bracis-33583 328 34 , , PUNCT bracis-33583 328 35 frank frank PROPN bracis-33583 328 36 , , PUNCT bracis-33583 328 37 e. e. PROPN bracis-33583 328 38 : : PUNCT bracis-33583 328 39 classifier classifier NOUN bracis-33583 328 40 chains chain NOUN bracis-33583 328 41 for for ADP bracis-33583 328 42 multi multi ADJ bracis-33583 328 43 - - ADJ bracis-33583 328 44 label label ADJ bracis-33583 328 45 classification classification NOUN bracis-33583 328 46 . . PUNCT bracis-33583 329 1 in in ADP bracis-33583 329 2 : : PUNCT bracis-33583 329 3 buntine buntine NOUN bracis-33583 329 4 , , PUNCT bracis-33583 329 5 w. w. PROPN bracis-33583 329 6 , , PUNCT bracis-33583 329 7 grobelnik grobelnik PROPN bracis-33583 329 8 , , PUNCT bracis-33583 329 9 m. m. NOUN bracis-33583 329 10 , , PUNCT bracis-33583 329 11 mladenić mladenić PROPN bracis-33583 329 12 , , PUNCT bracis-33583 329 13 d. d. PROPN bracis-33583 329 14 , , PUNCT bracis-33583 329 15 shawe shawe NOUN bracis-33583 329 16 - - PUNCT bracis-33583 329 17 taylor taylor PROPN bracis-33583 329 18 , , PUNCT bracis-33583 329 19 j. j. PROPN bracis-33583 329 20 ( ( PUNCT bracis-33583 329 21 eds eds PROPN bracis-33583 329 22 . . PUNCT bracis-33583 329 23 ) ) PUNCT bracis-33583 329 24 ecml ecml PROPN bracis-33583 329 25 pkdd pkdd NOUN bracis-33583 329 26 2009 2009 NUM bracis-33583 329 27 . . PUNCT bracis-33583 330 1 lncs lncs PROPN bracis-33583 330 2 ( ( PUNCT bracis-33583 330 3 lnai lnai ADJ bracis-33583 330 4 ) ) PUNCT bracis-33583 330 5 , , PUNCT bracis-33583 330 6 vol vol NOUN bracis-33583 330 7 . . PROPN bracis-33583 330 8 5782 5782 NUM bracis-33583 330 9 , , PUNCT bracis-33583 330 10 pp pp ADP bracis-33583 330 11 . . PUNCT bracis-33583 331 1 254–269 254–269 NUM bracis-33583 331 2 . . PUNCT bracis-33583 331 3 springer springer NOUN bracis-33583 331 4 , , PUNCT bracis-33583 331 5 heidelberg heidelberg PROPN bracis-33583 331 6 ( ( PUNCT bracis-33583 331 7 2009 2009 NUM bracis-33583 331 8 ) ) PUNCT bracis-33583 331 9 . . PUNCT bracis-33583 332 1 https://doi.org/10.1007/978-3-642-04174-7_17 https://doi.org/10.1007/978-3-642-04174-7_17 PROPN bracis-33583 332 2 chapter chapter NOUN bracis-33583 332 3     SPACE bracis-33583 332 4 math math NOUN bracis-33583 332 5     SPACE bracis-33583 332 6 google google PROPN bracis-33583 332 7 scholar scholar NOUN bracis-33583 332 8     SPACE bracis-33583 332 9 saini saini PROPN bracis-33583 332 10 , , PUNCT bracis-33583 332 11 n. n. PROPN bracis-33583 332 12 , , PUNCT bracis-33583 332 13 saha saha PROPN bracis-33583 332 14 , , PUNCT bracis-33583 332 15 s. s. PROPN bracis-33583 332 16 , , PUNCT bracis-33583 332 17 bhattacharyya bhattacharyya ADJ bracis-33583 332 18 , , PUNCT bracis-33583 332 19 p. p. NOUN bracis-33583 332 20 : : PUNCT bracis-33583 333 1 incorporation incorporation NOUN bracis-33583 333 2 of of ADP bracis-33583 333 3 neighborhood neighborhood NOUN bracis-33583 333 4 concept concept NOUN bracis-33583 333 5     SPACE bracis-33583 333 6 in in ADP bracis-33583 333 7 enhancing enhance VERB bracis-33583 333 8 som som PROPN bracis-33583 333 9 based base VERB bracis-33583 333 10 multi multi ADJ bracis-33583 333 11 - - ADJ bracis-33583 333 12 label label ADJ bracis-33583 333 13 classification classification NOUN bracis-33583 333 14 . . PUNCT bracis-33583 334 1 in in ADP bracis-33583 334 2 : : PUNCT bracis-33583 334 3 deka deka PROPN bracis-33583 334 4 , , PUNCT bracis-33583 334 5 b. b. PROPN bracis-33583 334 6 , , PUNCT bracis-33583 334 7 maji maji PROPN bracis-33583 334 8 , , PUNCT bracis-33583 334 9 p. p. PROPN bracis-33583 334 10 , , PUNCT bracis-33583 334 11 mitra mitra PROPN bracis-33583 334 12 , , PUNCT bracis-33583 334 13 s. s. PROPN bracis-33583 334 14 , , PUNCT bracis-33583 334 15 bhattacharyya bhattacharyya ADJ bracis-33583 334 16 , , PUNCT bracis-33583 334 17 d.k d.k PROPN bracis-33583 334 18 . . PROPN bracis-33583 334 19 , , PUNCT bracis-33583 334 20 bora bora PROPN bracis-33583 334 21 , , PUNCT bracis-33583 334 22 p.k p.k PROPN bracis-33583 334 23 . . PROPN bracis-33583 334 24 , , PUNCT bracis-33583 334 25 pal pal NOUN bracis-33583 334 26 , , PUNCT bracis-33583 334 27 s.k s.k PROPN bracis-33583 334 28 . . PUNCT bracis-33583 334 29 ( ( PUNCT bracis-33583 334 30 eds ed NOUN bracis-33583 334 31 . . PUNCT bracis-33583 334 32 ) ) PUNCT bracis-33583 334 33 premi premi PROPN bracis-33583 334 34 2019 2019 NUM bracis-33583 334 35 . . PUNCT bracis-33583 335 1 lncs lncs PROPN bracis-33583 335 2 , , PUNCT bracis-33583 335 3 vol vol NOUN bracis-33583 335 4 . . PROPN bracis-33583 335 5 11941 11941 NUM bracis-33583 335 6 , , PUNCT bracis-33583 335 7 pp pp ADJ bracis-33583 335 8 . . PUNCT bracis-33583 336 1 91–99 91–99 NUM bracis-33583 336 2 . . PUNCT bracis-33583 337 1 springer springer NOUN bracis-33583 337 2 , , PUNCT bracis-33583 337 3 cham cham PROPN bracis-33583 337 4 ( ( PUNCT bracis-33583 337 5 2019 2019 NUM bracis-33583 337 6 ) ) PUNCT bracis-33583 337 7 . . PUNCT bracis-33583 338 1 https://doi.org/10.1007/978-3-030-34869-4_11 https://doi.org/10.1007/978-3-030-34869-4_11 PROPN bracis-33583 338 2 chapter chapter NOUN bracis-33583 338 3     SPACE bracis-33583 338 4 math math NOUN bracis-33583 338 5     SPACE bracis-33583 338 6 google google PROPN bracis-33583 338 7 scholar scholar NOUN bracis-33583 338 8     SPACE bracis-33583 338 9 santos santo NOUN bracis-33583 338 10 , , PUNCT bracis-33583 338 11 b.z b.z PROPN bracis-33583 338 12 . . PROPN bracis-33583 338 13 , , PUNCT bracis-33583 338 14 nakano nakano PROPN bracis-33583 338 15 , , PUNCT bracis-33583 338 16 f.k f.k PROPN bracis-33583 338 17 . . PROPN bracis-33583 338 18 , , PUNCT bracis-33583 338 19 cerri cerri PROPN bracis-33583 338 20 , , PUNCT bracis-33583 338 21 r. r. PROPN bracis-33583 338 22 , , PUNCT bracis-33583 338 23 vens vens PROPN bracis-33583 338 24 , , PUNCT bracis-33583 338 25 c. c. NOUN bracis-33583 338 26 : : PUNCT bracis-33583 338 27 predictive predictive ADJ bracis-33583 338 28 bi bi ADJ bracis-33583 338 29 - - ADJ bracis-33583 338 30 clustering clustering ADJ bracis-33583 338 31 trees tree NOUN bracis-33583 338 32 for for ADP bracis-33583 338 33 hierarchical hierarchical ADJ bracis-33583 338 34 multi multi ADJ bracis-33583 338 35 - - ADJ bracis-33583 338 36 label label ADJ bracis-33583 338 37 classification classification NOUN bracis-33583 338 38 . . PUNCT bracis-33583 339 1 in in ADP bracis-33583 339 2 : : PUNCT bracis-33583 339 3 hutter hutter NOUN bracis-33583 339 4 , , PUNCT bracis-33583 339 5 f. f. PROPN bracis-33583 339 6 , , PUNCT bracis-33583 339 7 kersting kersting NOUN bracis-33583 339 8 , , PUNCT bracis-33583 339 9 k. k. PROPN bracis-33583 339 10 , , PUNCT bracis-33583 339 11 lijffijt lijffijt PROPN bracis-33583 339 12 , , PUNCT bracis-33583 339 13 j. j. PROPN bracis-33583 339 14 , , PUNCT bracis-33583 339 15 valera valera PROPN bracis-33583 339 16 , , PUNCT bracis-33583 339 17 i. i. PROPN bracis-33583 339 18 ( ( PUNCT bracis-33583 339 19 eds eds PROPN bracis-33583 339 20 . . PUNCT bracis-33583 339 21 ) ) PUNCT bracis-33583 339 22 ecml ecml PROPN bracis-33583 339 23 pkdd pkdd NOUN bracis-33583 339 24 2020 2020 NUM bracis-33583 339 25 . . PUNCT bracis-33583 340 1 lncs lncs PROPN bracis-33583 340 2 ( ( PUNCT bracis-33583 340 3 lnai lnai ADJ bracis-33583 340 4 ) ) PUNCT bracis-33583 340 5 , , PUNCT bracis-33583 340 6 vol vol NOUN bracis-33583 340 7 . . PROPN bracis-33583 340 8 12459 12459 NUM bracis-33583 340 9 , , PUNCT bracis-33583 340 10 pp pp ADJ bracis-33583 340 11 . . PUNCT bracis-33583 341 1 701–718 701–718 NUM bracis-33583 341 2 . . PUNCT bracis-33583 341 3 springer springer NOUN bracis-33583 341 4 , , PUNCT bracis-33583 341 5 cham cham PROPN bracis-33583 341 6 ( ( PUNCT bracis-33583 341 7 2021 2021 NUM bracis-33583 341 8 ) ) PUNCT bracis-33583 341 9 . . PUNCT bracis-33583 342 1 https://doi.org/10.1007/978-3-030-67664-3_42 https://doi.org/10.1007/978-3-030-67664-3_42 PROPN bracis-33583 342 2 chapter chapter NOUN bracis-33583 342 3     SPACE bracis-33583 342 4 math math NOUN bracis-33583 342 5     SPACE bracis-33583 342 6 google google PROPN bracis-33583 342 7 scholar scholar NOUN bracis-33583 342 8     SPACE bracis-33583 342 9 schietgat schietgat PROPN bracis-33583 342 10 , , PUNCT bracis-33583 342 11 l. l. PROPN bracis-33583 342 12 , , PUNCT bracis-33583 342 13 vens vens PROPN bracis-33583 342 14 , , PUNCT bracis-33583 342 15 c. c. PROPN bracis-33583 342 16 , , PUNCT bracis-33583 342 17 struyf struyf PROPN bracis-33583 342 18 , , PUNCT bracis-33583 342 19 j. j. PROPN bracis-33583 342 20 , , PUNCT bracis-33583 342 21 blockeel blockeel PROPN bracis-33583 342 22 , , PUNCT bracis-33583 342 23 h. h. PROPN bracis-33583 342 24 , , PUNCT bracis-33583 342 25 kocev kocev PROPN bracis-33583 342 26 , , PUNCT bracis-33583 342 27 d. d. PROPN bracis-33583 342 28 , , PUNCT bracis-33583 342 29 džeroski džeroski PROPN bracis-33583 342 30 , , PUNCT bracis-33583 342 31 s. s. PROPN bracis-33583 342 32 : : PUNCT bracis-33583 342 33 predicting predict VERB bracis-33583 342 34 gene gene NOUN bracis-33583 342 35 function function NOUN bracis-33583 342 36 using use VERB bracis-33583 342 37 hierarchical hierarchical ADJ bracis-33583 342 38 multi multi ADJ bracis-33583 342 39 - - ADJ bracis-33583 342 40 label label ADJ bracis-33583 342 41 decision decision NOUN bracis-33583 342 42 tree tree NOUN bracis-33583 342 43 ensembles ensemble NOUN bracis-33583 342 44 . . PUNCT bracis-33583 343 1 bmc bmc PROPN bracis-33583 343 2 bioinf bioinf PROPN bracis-33583 343 3 . . PUNCT bracis-33583 344 1 11(1 11(1 NUM bracis-33583 344 2 ) ) PUNCT bracis-33583 344 3 , , PUNCT bracis-33583 344 4 1–14 1–14 PROPN bracis-33583 344 5 ( ( PUNCT bracis-33583 344 6 2010 2010 NUM bracis-33583 344 7 ) ) PUNCT bracis-33583 344 8 article article NOUN bracis-33583 344 9     SPACE bracis-33583 344 10 math math NOUN bracis-33583 344 11     SPACE bracis-33583 344 12 google google PROPN bracis-33583 344 13 scholar scholar NOUN bracis-33583 344 14     SPACE bracis-33583 344 15 sechidis sechidi NOUN bracis-33583 344 16 , , PUNCT bracis-33583 344 17 k. k. PROPN bracis-33583 344 18 , , PUNCT bracis-33583 344 19 tsoumakas tsoumakas PROPN bracis-33583 344 20 , , PUNCT bracis-33583 344 21 g. g. PROPN bracis-33583 344 22 , , PUNCT bracis-33583 344 23 vlahavas vlahavas PROPN bracis-33583 344 24 , , PUNCT bracis-33583 344 25 i. i. NOUN bracis-33583 344 26 : : PUNCT bracis-33583 344 27 on on ADP bracis-33583 344 28 the the DET bracis-33583 344 29 stratification stratification NOUN bracis-33583 344 30 of of ADP bracis-33583 344 31 multi multi ADJ bracis-33583 344 32 - - ADJ bracis-33583 344 33 label label ADJ bracis-33583 344 34 data datum NOUN bracis-33583 344 35 . . PUNCT bracis-33583 345 1 in in ADP bracis-33583 345 2 : : PUNCT bracis-33583 345 3 gunopulos gunopulo NOUN bracis-33583 345 4 , , PUNCT bracis-33583 345 5 d. d. PROPN bracis-33583 345 6 , , PUNCT bracis-33583 345 7 hofmann hofmann PROPN bracis-33583 345 8 , , PUNCT bracis-33583 345 9 t. t. PROPN bracis-33583 345 10 , , PUNCT bracis-33583 345 11 malerba malerba PROPN bracis-33583 345 12 , , PUNCT bracis-33583 345 13 d. d. PROPN bracis-33583 345 14 , , PUNCT bracis-33583 345 15 vazirgiannis vazirgiannis PROPN bracis-33583 345 16 , , PUNCT bracis-33583 345 17 m. m. NOUN bracis-33583 345 18 ( ( PUNCT bracis-33583 345 19 eds ed NOUN bracis-33583 345 20 . . PUNCT bracis-33583 345 21 ) ) PUNCT bracis-33583 346 1 ecml ecml PROPN bracis-33583 346 2 pkdd pkdd NOUN bracis-33583 346 3 2011 2011 NUM bracis-33583 346 4 , , PUNCT bracis-33583 346 5 part part NOUN bracis-33583 346 6 iii iii NOUN bracis-33583 346 7 . . PROPN bracis-33583 346 8 lncs lncs PROPN bracis-33583 346 9 ( ( PUNCT bracis-33583 346 10 lnai lnai ADJ bracis-33583 346 11 ) ) PUNCT bracis-33583 346 12 , , PUNCT bracis-33583 346 13 vol vol NOUN bracis-33583 346 14 . . PROPN bracis-33583 346 15 6913 6913 NUM bracis-33583 346 16 , , PUNCT bracis-33583 346 17 pp pp ADV bracis-33583 346 18 . . PUNCT bracis-33583 347 1 145–158 145–158 NUM bracis-33583 347 2 . . PUNCT bracis-33583 347 3 springer springer NOUN bracis-33583 347 4 , , PUNCT bracis-33583 347 5 heidelberg heidelberg PROPN bracis-33583 347 6 ( ( PUNCT bracis-33583 347 7 2011 2011 NUM bracis-33583 347 8 ) ) PUNCT bracis-33583 347 9 . . PUNCT bracis-33583 348 1 https://doi.org/10.1007/978-3-642-23808-6_10 https://doi.org/10.1007/978-3-642-23808-6_10 PROPN bracis-33583 348 2 chapter chapter NOUN bracis-33583 348 3     SPACE bracis-33583 348 4 math math NOUN bracis-33583 348 5     SPACE bracis-33583 348 6 google google PROPN bracis-33583 348 7 scholar scholar NOUN bracis-33583 348 8     SPACE bracis-33583 348 9 spyromitros spyromitro NOUN bracis-33583 348 10 , , PUNCT bracis-33583 348 11 e. e. PROPN bracis-33583 348 12 , , PUNCT bracis-33583 348 13 tsoumakas tsoumakas PROPN bracis-33583 348 14 , , PUNCT bracis-33583 348 15 g. g. PROPN bracis-33583 348 16 , , PUNCT bracis-33583 348 17 vlahavas vlahavas PROPN bracis-33583 348 18 , , PUNCT bracis-33583 348 19 i. i. NOUN bracis-33583 348 20 : : PUNCT bracis-33583 348 21 an an DET bracis-33583 348 22 empirical empirical ADJ bracis-33583 348 23 study study NOUN bracis-33583 348 24 of of ADP bracis-33583 348 25 lazy lazy ADJ bracis-33583 348 26 multilabel multilabel NOUN bracis-33583 348 27 classification classification NOUN bracis-33583 348 28 algorithms algorithm NOUN bracis-33583 348 29 . . PUNCT bracis-33583 349 1 in in ADP bracis-33583 349 2 : : PUNCT bracis-33583 349 3 darzentas darzenta NOUN bracis-33583 349 4 , , PUNCT bracis-33583 349 5 j. j. PROPN bracis-33583 349 6 , , PUNCT bracis-33583 349 7 vouros vouros PROPN bracis-33583 349 8 , , PUNCT bracis-33583 349 9 g.a g.a PROPN bracis-33583 349 10 . . PROPN bracis-33583 349 11 , , PUNCT bracis-33583 349 12 vosinakis vosinakis PROPN bracis-33583 349 13 , , PUNCT bracis-33583 349 14 s. s. PROPN bracis-33583 349 15 , , PUNCT bracis-33583 349 16 arnellos arnellos PROPN bracis-33583 349 17 , , PUNCT bracis-33583 349 18 a. a. NOUN bracis-33583 349 19 ( ( PUNCT bracis-33583 349 20 eds eds PROPN bracis-33583 349 21 . . PUNCT bracis-33583 349 22 ) ) PUNCT bracis-33583 349 23 setn setn NOUN bracis-33583 349 24 2008 2008 NUM bracis-33583 349 25 . . PUNCT bracis-33583 350 1 lncs lncs PROPN bracis-33583 350 2 ( ( PUNCT bracis-33583 350 3 lnai lnai ADJ bracis-33583 350 4 ) ) PUNCT bracis-33583 350 5 , , PUNCT bracis-33583 350 6 vol vol NOUN bracis-33583 350 7 . . PUNCT bracis-33583 350 8 5138 5138 NUM bracis-33583 350 9 , , PUNCT bracis-33583 350 10 pp pp ADP bracis-33583 350 11 . . PUNCT bracis-33583 351 1 401–406 401–406 NUM bracis-33583 351 2 . . PUNCT bracis-33583 351 3 springer springer NOUN bracis-33583 351 4 , , PUNCT bracis-33583 351 5 heidelberg heidelberg PROPN bracis-33583 351 6 ( ( PUNCT bracis-33583 351 7 2008 2008 NUM bracis-33583 351 8 ) ) PUNCT bracis-33583 351 9 . . PUNCT bracis-33583 352 1 https://doi.org/10.1007/978-3-540-87881-0_40 https://doi.org/10.1007/978-3-540-87881-0_40 PUNCT bracis-33583 352 2 chapter chapter NOUN bracis-33583 352 3     SPACE bracis-33583 352 4 math math NOUN bracis-33583 352 5     SPACE bracis-33583 352 6 google google PROPN bracis-33583 352 7 scholar scholar NOUN bracis-33583 352 8     SPACE bracis-33583 352 9 trohidis trohidi NOUN bracis-33583 352 10 , , PUNCT bracis-33583 352 11 k. k. PROPN bracis-33583 352 12 , , PUNCT bracis-33583 352 13 tsoumakas tsoumakas PROPN bracis-33583 352 14 , , PUNCT bracis-33583 352 15 g. g. PROPN bracis-33583 352 16 , , PUNCT bracis-33583 352 17 kalliris kalliris PROPN bracis-33583 352 18 , , PUNCT bracis-33583 352 19 g. g. PROPN bracis-33583 352 20 , , PUNCT bracis-33583 352 21 vlahavas vlahavas PROPN bracis-33583 352 22 , , PUNCT bracis-33583 352 23 i.p i.p PROPN bracis-33583 352 24 . . PROPN bracis-33583 352 25 , , PUNCT bracis-33583 352 26 et et PROPN bracis-33583 352 27     SPACE bracis-33583 352 28 al al PROPN bracis-33583 352 29 . . PROPN bracis-33583 352 30 : : PUNCT bracis-33583 353 1 multi multi ADJ bracis-33583 353 2 - - ADJ bracis-33583 353 3 label label ADJ bracis-33583 353 4 classification classification NOUN bracis-33583 353 5 of of ADP bracis-33583 353 6 music music NOUN bracis-33583 353 7 into into ADP bracis-33583 353 8 emotions emotion NOUN bracis-33583 353 9 . . PUNCT bracis-33583 354 1 in in ADP bracis-33583 354 2 : : PUNCT bracis-33583 354 3 ismir ismir ADJ bracis-33583 354 4 , , PUNCT bracis-33583 354 5 vol vol NOUN bracis-33583 354 6 . . PUNCT bracis-33583 354 7     SPACE bracis-33583 354 8 8 8 NUM bracis-33583 354 9 , , PUNCT bracis-33583 354 10 pp pp ADJ bracis-33583 354 11 . . PUNCT bracis-33583 355 1 325–330 325–330 NUM bracis-33583 355 2 ( ( PUNCT bracis-33583 355 3 2008 2008 NUM bracis-33583 355 4 ) ) PUNCT bracis-33583 355 5 google google PROPN bracis-33583 355 6 scholar scholar NOUN bracis-33583 355 7     SPACE bracis-33583 355 8 tsoumakas tsoumakas NOUN bracis-33583 355 9 , , PUNCT bracis-33583 355 10 g. g. PROPN bracis-33583 355 11 , , PUNCT bracis-33583 355 12 katakis katakis PROPN bracis-33583 355 13 , , PUNCT bracis-33583 355 14 i. i. PROPN bracis-33583 355 15 , , PUNCT bracis-33583 355 16 vlahavas vlahavas PROPN bracis-33583 355 17 , , PUNCT bracis-33583 355 18 i. i. NOUN bracis-33583 355 19 : : PUNCT bracis-33583 355 20 effective effective ADJ bracis-33583 355 21 and and CCONJ bracis-33583 355 22 efficient efficient ADJ bracis-33583 355 23 multilabel multilabel NOUN bracis-33583 355 24 classification classification NOUN bracis-33583 355 25 in in ADP bracis-33583 355 26 domains domain NOUN bracis-33583 355 27 with with ADP bracis-33583 355 28 large large ADJ bracis-33583 355 29 number number NOUN bracis-33583 355 30 of of ADP bracis-33583 355 31 labels label NOUN bracis-33583 355 32 . . PUNCT bracis-33583 356 1 in in ADP bracis-33583 356 2 : : PUNCT bracis-33583 356 3 proceeding proceeding NOUN bracis-33583 356 4 of of ADP bracis-33583 356 5 the the DET bracis-33583 356 6 ecml ecml NOUN bracis-33583 356 7 / / SYM bracis-33583 356 8 pkdd pkdd NOUN bracis-33583 356 9 2008 2008 NUM bracis-33583 356 10 workshop workshop NOUN bracis-33583 356 11 on on ADP bracis-33583 356 12 mining mining NOUN bracis-33583 356 13 multidimensional multidimensional ADJ bracis-33583 356 14 data datum NOUN bracis-33583 356 15 , , PUNCT bracis-33583 356 16 mmd mmd PROPN bracis-33583 356 17 2008 2008 NUM bracis-33583 356 18 , , PUNCT bracis-33583 356 19 vol vol NOUN bracis-33583 356 20 . . PUNCT bracis-33583 356 21     SPACE bracis-33583 356 22 21 21 NUM bracis-33583 356 23 , , PUNCT bracis-33583 356 24 pp pp ADJ bracis-33583 356 25 . . PUNCT bracis-33583 357 1 53–59 53–59 NUM bracis-33583 357 2 ( ( PUNCT bracis-33583 357 3 2008 2008 NUM bracis-33583 357 4 ) ) PUNCT bracis-33583 357 5 google google PROPN bracis-33583 357 6 scholar scholar NOUN bracis-33583 357 7     SPACE bracis-33583 357 8 tsoumakas tsoumakas NOUN bracis-33583 357 9 , , PUNCT bracis-33583 357 10 g. g. PROPN bracis-33583 357 11 , , PUNCT bracis-33583 357 12 katakis katakis PROPN bracis-33583 357 13 , , PUNCT bracis-33583 357 14 i. i. PROPN bracis-33583 357 15 , , PUNCT bracis-33583 357 16 vlahavas vlahavas PROPN bracis-33583 357 17 , , PUNCT bracis-33583 357 18 i. i. NOUN bracis-33583 357 19 : : PUNCT bracis-33583 357 20 mining mine VERB bracis-33583 357 21 multi multi ADJ bracis-33583 357 22 - - ADJ bracis-33583 357 23 label label ADJ bracis-33583 357 24 data datum NOUN bracis-33583 357 25 . . PUNCT bracis-33583 358 1 in in ADP bracis-33583 358 2 : : PUNCT bracis-33583 358 3 data datum NOUN bracis-33583 358 4 mining mining NOUN bracis-33583 358 5 and and CCONJ bracis-33583 358 6 knowledge knowledge NOUN bracis-33583 358 7 discovery discovery PROPN bracis-33583 358 8 handbook handbook NOUN bracis-33583 358 9 , , PUNCT bracis-33583 358 10 pp pp ADP bracis-33583 358 11 . . PUNCT bracis-33583 359 1 667–685 667–685 NUM bracis-33583 359 2 ( ( PUNCT bracis-33583 359 3 2010 2010 NUM bracis-33583 359 4 ) ) PUNCT bracis-33583 359 5 google google PROPN bracis-33583 359 6 scholar scholar NOUN bracis-33583 359 7     SPACE bracis-33583 359 8 tsoumakas tsoumakas NOUN bracis-33583 359 9 , , PUNCT bracis-33583 359 10 g. g. PROPN bracis-33583 359 11 , , PUNCT bracis-33583 359 12 spyromitros spyromitro NOUN bracis-33583 359 13 - - PUNCT bracis-33583 359 14 xioufis xioufis PROPN bracis-33583 359 15 , , PUNCT bracis-33583 359 16 e. e. PROPN bracis-33583 359 17 , , PUNCT bracis-33583 359 18 vilcek vilcek PROPN bracis-33583 359 19 , , PUNCT bracis-33583 359 20 j. j. PROPN bracis-33583 359 21 , , PUNCT bracis-33583 359 22 vlahavas vlahavas PROPN bracis-33583 359 23 , , PUNCT bracis-33583 359 24 i. i. NOUN bracis-33583 359 25 : : PUNCT bracis-33583 359 26 mulan mulan PROPN bracis-33583 359 27 : : PUNCT bracis-33583 359 28 a a DET bracis-33583 359 29 java java NOUN bracis-33583 359 30 library library NOUN bracis-33583 359 31 for for ADP bracis-33583 359 32 multi multi ADJ bracis-33583 359 33 - - ADJ bracis-33583 359 34 label label ADJ bracis-33583 359 35 learning learning NOUN bracis-33583 359 36 . . PUNCT bracis-33583 360 1 j. j. PROPN bracis-33583 360 2 mach mach PROPN bracis-33583 360 3 . . PUNCT bracis-33583 361 1 learn learn VERB bracis-33583 361 2 . . PUNCT bracis-33583 362 1 res re NOUN bracis-33583 362 2 . . PROPN bracis-33583 363 1 12 12 NUM bracis-33583 363 2 , , PUNCT bracis-33583 363 3 2411–2414 2411–2414 NUM bracis-33583 363 4 ( ( PUNCT bracis-33583 363 5 2011 2011 NUM bracis-33583 363 6 ) ) PUNCT bracis-33583 363 7 mathscinet mathscinet NOUN bracis-33583 363 8     SPACE bracis-33583 363 9 math math NOUN bracis-33583 363 10     SPACE bracis-33583 363 11 google google PROPN bracis-33583 363 12 scholar scholar NOUN bracis-33583 363 13     SPACE bracis-33583 363 14 tsoumakas tsoumakas NOUN bracis-33583 363 15 , , PUNCT bracis-33583 363 16 g. g. PROPN bracis-33583 363 17 , , PUNCT bracis-33583 363 18 vlahavas vlahavas PROPN bracis-33583 363 19 , , PUNCT bracis-33583 363 20 i. i. NOUN bracis-33583 363 21 : : PUNCT bracis-33583 363 22 random random ADJ bracis-33583 363 23 k k NOUN bracis-33583 363 24 - - PUNCT bracis-33583 363 25 labelsets labelset NOUN bracis-33583 363 26 : : PUNCT bracis-33583 363 27 an an DET bracis-33583 363 28 ensemble ensemble ADJ bracis-33583 363 29 method method NOUN bracis-33583 363 30 for for ADP bracis-33583 363 31 multilabel multilabel NOUN bracis-33583 363 32 classification classification NOUN bracis-33583 363 33 . . PUNCT bracis-33583 364 1 in in ADP bracis-33583 364 2 : : PUNCT bracis-33583 364 3 kok kok PROPN bracis-33583 364 4 , , PUNCT bracis-33583 364 5 j.n j.n PROPN bracis-33583 364 6 . . PROPN bracis-33583 364 7 , , PUNCT bracis-33583 364 8 koronacki koronacki PROPN bracis-33583 364 9 , , PUNCT bracis-33583 364 10 j. j. PROPN bracis-33583 364 11 , , PUNCT bracis-33583 364 12 mantaras mantaras PROPN bracis-33583 364 13 , , PUNCT bracis-33583 364 14 r.l r.l PROPN bracis-33583 364 15 . . PROPN bracis-33583 364 16 , , PUNCT bracis-33583 364 17 matwin matwin PROPN bracis-33583 364 18 , , PUNCT bracis-33583 364 19 s. s. PROPN bracis-33583 364 20 , , PUNCT bracis-33583 364 21 mladenič mladenič PROPN bracis-33583 364 22 , , PUNCT bracis-33583 364 23 d. d. PROPN bracis-33583 364 24 , , PUNCT bracis-33583 364 25 skowron skowron PROPN bracis-33583 364 26 , , PUNCT bracis-33583 364 27 a. a. NOUN bracis-33583 364 28 ( ( PUNCT bracis-33583 364 29 eds eds PROPN bracis-33583 364 30 . . PUNCT bracis-33583 364 31 ) ) PUNCT bracis-33583 364 32 ecml ecml PROPN bracis-33583 364 33 2007 2007 NUM bracis-33583 364 34 . . PUNCT bracis-33583 365 1 lncs lncs PROPN bracis-33583 365 2 ( ( PUNCT bracis-33583 365 3 lnai lnai ADJ bracis-33583 365 4 ) ) PUNCT bracis-33583 365 5 , , PUNCT bracis-33583 365 6 vol vol NOUN bracis-33583 365 7 . . PROPN bracis-33583 365 8 4701 4701 NUM bracis-33583 365 9 , , PUNCT bracis-33583 365 10 pp pp ADJ bracis-33583 365 11 . . PUNCT bracis-33583 366 1 406–417 406–417 NUM bracis-33583 366 2 . . PUNCT bracis-33583 366 3 springer springer NOUN bracis-33583 366 4 , , PUNCT bracis-33583 366 5 heidelberg heidelberg PROPN bracis-33583 366 6 ( ( PUNCT bracis-33583 366 7 2007 2007 NUM bracis-33583 366 8 ) ) PUNCT bracis-33583 366 9 . . PUNCT bracis-33583 367 1 https://doi.org/10.1007/978-3-540-74958-5_38 https://doi.org/10.1007/978-3-540-74958-5_38 PROPN bracis-33583 367 2 chapter chapter NOUN bracis-33583 367 3     SPACE bracis-33583 367 4 math math NOUN bracis-33583 367 5     SPACE bracis-33583 367 6 google google PROPN bracis-33583 367 7 scholar scholar NOUN bracis-33583 367 8     SPACE bracis-33583 367 9 vapnik vapnik NOUN bracis-33583 367 10 , , PUNCT bracis-33583 367 11 v.n v.n PROPN bracis-33583 367 12 . . PROPN bracis-33583 367 13 : : PUNCT bracis-33583 368 1 the the DET bracis-33583 368 2 nature nature NOUN bracis-33583 368 3 of of ADP bracis-33583 368 4 statistical statistical ADJ bracis-33583 368 5 learning learning NOUN bracis-33583 368 6 theory theory NOUN bracis-33583 368 7 . . PUNCT bracis-33583 369 1 information information NOUN bracis-33583 369 2 science science NOUN bracis-33583 369 3 and and CCONJ bracis-33583 369 4 statistics statistic NOUN bracis-33583 369 5 , , PUNCT bracis-33583 369 6 springer springer NOUN bracis-33583 369 7 , , PUNCT bracis-33583 369 8 new new PROPN bracis-33583 369 9 york york PROPN bracis-33583 369 10 ( ( PUNCT bracis-33583 369 11 1999 1999 NUM bracis-33583 369 12 ) ) PUNCT bracis-33583 369 13 . . PUNCT bracis-33583 370 1 https://doi.org/10.1007/978-1-4757-3264-1 https://doi.org/10.1007/978-1-4757-3264-1 ADJ bracis-33583 370 2 book book NOUN bracis-33583 370 3     SPACE bracis-33583 370 4 math math NOUN bracis-33583 370 5     SPACE bracis-33583 370 6 google google PROPN bracis-33583 370 7 scholar scholar NOUN bracis-33583 370 8     SPACE bracis-33583 370 9 vens vens PROPN bracis-33583 370 10 , , PUNCT bracis-33583 370 11 c. c. PROPN bracis-33583 370 12 , , PUNCT bracis-33583 370 13 struyf struyf PROPN bracis-33583 370 14 , , PUNCT bracis-33583 370 15 j. j. PROPN bracis-33583 370 16 , , PUNCT bracis-33583 370 17 schietgat schietgat PROPN bracis-33583 370 18 , , PUNCT bracis-33583 370 19 l. l. PROPN bracis-33583 370 20 , , PUNCT bracis-33583 370 21 džeroski džeroski PROPN bracis-33583 370 22 , , PUNCT bracis-33583 370 23 s. s. PROPN bracis-33583 370 24 , , PUNCT bracis-33583 370 25 blockeel blockeel PROPN bracis-33583 370 26 , , PUNCT bracis-33583 370 27 h. h. NOUN bracis-33583 370 28 : : PUNCT bracis-33583 370 29 decision decision NOUN bracis-33583 370 30 trees tree NOUN bracis-33583 370 31 for for ADP bracis-33583 370 32 hierarchical hierarchical ADJ bracis-33583 370 33 multi multi ADJ bracis-33583 370 34 - - ADJ bracis-33583 370 35 label label ADJ bracis-33583 370 36 classification classification NOUN bracis-33583 370 37 . . PUNCT bracis-33583 371 1 mach mach NOUN bracis-33583 371 2 . . PUNCT bracis-33583 372 1 learn learn VERB bracis-33583 372 2 . . PUNCT bracis-33583 373 1 73(2 73(2 NUM bracis-33583 373 2 ) ) PUNCT bracis-33583 373 3 , , PUNCT bracis-33583 373 4 185–214 185–214 NUM bracis-33583 373 5 ( ( PUNCT bracis-33583 373 6 2008 2008 NUM bracis-33583 373 7 ) ) PUNCT bracis-33583 373 8 article article NOUN bracis-33583 373 9     SPACE bracis-33583 373 10 math math NOUN bracis-33583 373 11     SPACE bracis-33583 373 12 google google PROPN bracis-33583 373 13 scholar scholar NOUN bracis-33583 373 14     SPACE bracis-33583 373 15 wu wu PROPN bracis-33583 373 16 , , PUNCT bracis-33583 373 17 b. b. PROPN bracis-33583 373 18 , , PUNCT bracis-33583 373 19 zhong zhong PROPN bracis-33583 373 20 , , PUNCT bracis-33583 373 21 e. e. PROPN bracis-33583 373 22 , , PUNCT bracis-33583 373 23 horner horner PROPN bracis-33583 373 24 , , PUNCT bracis-33583 373 25 a. a. PROPN bracis-33583 373 26 , , PUNCT bracis-33583 373 27 yang yang PROPN bracis-33583 373 28 , , PUNCT bracis-33583 373 29 q. q. PROPN bracis-33583 373 30 : : PUNCT bracis-33583 373 31 music music NOUN bracis-33583 373 32 emotion emotion NOUN bracis-33583 373 33 recognition recognition NOUN bracis-33583 373 34 by by ADP bracis-33583 373 35 multi multi ADJ bracis-33583 373 36 - - ADJ bracis-33583 373 37 label label ADJ bracis-33583 373 38 multi multi ADJ bracis-33583 373 39 - - ADJ bracis-33583 373 40 layer layer ADJ bracis-33583 373 41 multi multi ADJ bracis-33583 373 42 - - ADJ bracis-33583 373 43 instance instance ADJ bracis-33583 373 44 multi multi ADJ bracis-33583 373 45 - - ADJ bracis-33583 373 46 view view ADJ bracis-33583 373 47 learning learning NOUN bracis-33583 373 48 . . PUNCT bracis-33583 374 1 in in ADP bracis-33583 374 2 : : PUNCT bracis-33583 374 3 proceedings proceeding NOUN bracis-33583 374 4 of of ADP bracis-33583 374 5 the the DET bracis-33583 374 6 22nd 22nd PROPN bracis-33583 374 7 acm acm PROPN bracis-33583 374 8 international international ADJ bracis-33583 374 9 conference conference NOUN bracis-33583 374 10 on on ADP bracis-33583 374 11 multimedia multimedia NOUN bracis-33583 374 12 , , PUNCT bracis-33583 374 13 pp pp ADJ bracis-33583 374 14 . . PUNCT bracis-33583 375 1 117–126 117–126 NUM bracis-33583 375 2 ( ( PUNCT bracis-33583 375 3 2014 2014 NUM bracis-33583 375 4 ) ) PUNCT bracis-33583 375 5 google google PROPN bracis-33583 375 6 scholar scholar NOUN bracis-33583 375 7     SPACE bracis-33583 375 8 zhang zhang PROPN bracis-33583 375 9 , , PUNCT bracis-33583 375 10 m.l m.l PROPN bracis-33583 375 11 . . PROPN bracis-33583 375 12 , , PUNCT bracis-33583 375 13 zhou zhou PROPN bracis-33583 375 14 , , PUNCT bracis-33583 375 15 z.h z.h PROPN bracis-33583 375 16 . . PROPN bracis-33583 375 17 : : PUNCT bracis-33583 375 18 a a DET bracis-33583 375 19 k k X bracis-33583 375 20 - - PUNCT bracis-33583 375 21 nearest near ADJ bracis-33583 375 22 neighbor neighbor NOUN bracis-33583 375 23 based base VERB bracis-33583 375 24 algorithm algorithm NOUN bracis-33583 375 25 for for ADP bracis-33583 375 26 multi multi ADJ bracis-33583 375 27 - - ADJ bracis-33583 375 28 label label ADJ bracis-33583 375 29 classification classification NOUN bracis-33583 375 30 . . PUNCT bracis-33583 376 1 in in ADP bracis-33583 376 2 : : PUNCT bracis-33583 376 3 2005 2005 NUM bracis-33583 376 4 ieee ieee PROPN bracis-33583 376 5 international international ADJ bracis-33583 376 6 conference conference NOUN bracis-33583 376 7 on on ADP bracis-33583 376 8 granular granular ADJ bracis-33583 376 9 computing computing NOUN bracis-33583 376 10 , , PUNCT bracis-33583 376 11 vol vol NOUN bracis-33583 376 12 . . PUNCT bracis-33583 376 13     SPACE bracis-33583 376 14 2 2 NUM bracis-33583 376 15 , , PUNCT bracis-33583 376 16 pp pp ADJ bracis-33583 376 17 . . PUNCT bracis-33583 377 1 718–721 718–721 NUM bracis-33583 377 2 . . PUNCT bracis-33583 378 1 ieee ieee NOUN bracis-33583 378 2 ( ( PUNCT bracis-33583 378 3 2005 2005 NUM bracis-33583 378 4 ) ) PUNCT bracis-33583 378 5 google google PROPN bracis-33583 378 6 scholar scholar NOUN bracis-33583 378 7     SPACE bracis-33583 378 8 zhang zhang PROPN bracis-33583 378 9 , , PUNCT bracis-33583 378 10 m.l m.l PROPN bracis-33583 378 11 . . PROPN bracis-33583 378 12 , , PUNCT bracis-33583 378 13 zhou zhou PROPN bracis-33583 378 14 , , PUNCT bracis-33583 378 15 z.h z.h PROPN bracis-33583 378 16 . . PROPN bracis-33583 378 17 : : PUNCT bracis-33583 378 18 multilabel multilabel NOUN bracis-33583 378 19 neural neural ADJ bracis-33583 378 20 networks network NOUN bracis-33583 378 21 with with ADP bracis-33583 378 22 applications application NOUN bracis-33583 378 23 to to ADP bracis-33583 378 24 functional functional ADJ bracis-33583 378 25 genomics genomic NOUN bracis-33583 378 26 and and CCONJ bracis-33583 378 27 text text NOUN bracis-33583 378 28 categorization categorization NOUN bracis-33583 378 29 . . PUNCT bracis-33583 379 1 ieee ieee PROPN bracis-33583 379 2 trans trans PROPN bracis-33583 379 3 . . PROPN bracis-33583 380 1 knowl knowl PROPN bracis-33583 380 2 . . PUNCT bracis-33583 381 1 data data PROPN bracis-33583 381 2 eng eng PROPN bracis-33583 381 3 . . PROPN bracis-33583 381 4 18(10 18(10 NUM bracis-33583 381 5 ) ) PUNCT bracis-33583 381 6 , , PUNCT bracis-33583 381 7 1338–1351 1338–1351 NUM bracis-33583 381 8 ( ( PUNCT bracis-33583 381 9 2006 2006 NUM bracis-33583 381 10 ) ) PUNCT bracis-33583 381 11 article article NOUN bracis-33583 381 12     SPACE bracis-33583 381 13 math math NOUN bracis-33583 381 14     SPACE bracis-33583 381 15 google google PROPN bracis-33583 381 16 scholar scholar NOUN bracis-33583 381 17     SPACE bracis-33583 381 18 zhang zhang PROPN bracis-33583 381 19 , , PUNCT bracis-33583 381 20 m.l m.l PROPN bracis-33583 381 21 . . PROPN bracis-33583 381 22 , , PUNCT bracis-33583 381 23 zhou zhou PROPN bracis-33583 381 24 , , PUNCT bracis-33583 381 25 z.h z.h PROPN bracis-33583 381 26 . . PROPN bracis-33583 381 27 : : PUNCT bracis-33583 381 28 ml ml PROPN bracis-33583 381 29 - - PUNCT bracis-33583 381 30 knn knn VERB bracis-33583 381 31 : : PUNCT bracis-33583 381 32 a a DET bracis-33583 381 33 lazy lazy ADJ bracis-33583 381 34 learning learning NOUN bracis-33583 381 35 approach approach NOUN bracis-33583 381 36 to to ADP bracis-33583 381 37 multi multi ADJ bracis-33583 381 38 - - ADJ bracis-33583 381 39 label label ADJ bracis-33583 381 40 learning learning NOUN bracis-33583 381 41 . . PUNCT bracis-33583 382 1 pattern pattern NOUN bracis-33583 382 2 recogn recogn VERB bracis-33583 382 3 . . PUNCT bracis-33583 383 1 40(7 40(7 NOUN bracis-33583 383 2 ) ) PUNCT bracis-33583 383 3 , , PUNCT bracis-33583 383 4 2038–2048 2038–2048 NUM bracis-33583 383 5 ( ( PUNCT bracis-33583 383 6 2007 2007 NUM bracis-33583 383 7 ) ) PUNCT bracis-33583 383 8 article article NOUN bracis-33583 383 9     SPACE bracis-33583 383 10 math math NOUN bracis-33583 383 11     SPACE bracis-33583 383 12 google google PROPN bracis-33583 383 13 scholar scholar NOUN bracis-33583 383 14     SPACE bracis-33583 383 15 download download NOUN bracis-33583 383 16 references reference NOUN bracis-33583 383 17 acknowledgment acknowledgment NOUN bracis-33583 383 18 this this DET bracis-33583 383 19 study study NOUN bracis-33583 383 20 was be AUX bracis-33583 383 21 financed finance VERB bracis-33583 383 22 by by ADP bracis-33583 383 23 the the DET bracis-33583 383 24 national national PROPN bracis-33583 383 25 council council PROPN bracis-33583 383 26 for for ADP bracis-33583 383 27 scientific scientific ADJ bracis-33583 383 28 and and CCONJ bracis-33583 383 29 technological technological ADJ bracis-33583 383 30 development development NOUN bracis-33583 383 31 ( ( PUNCT bracis-33583 383 32 cnpq cnpq NOUN bracis-33583 383 33 ) ) PUNCT bracis-33583 383 34 and and CCONJ bracis-33583 384 1 the the DET bracis-33583 384 2 são são PROPN bracis-33583 384 3 paulo paulo PROPN bracis-33583 384 4 research research PROPN bracis-33583 384 5 foundation foundation PROPN bracis-33583 384 6 ( ( PUNCT bracis-33583 384 7 fapesp fapesp NOUN bracis-33583 384 8 ) ) PUNCT bracis-33583 384 9 grant grant VERB bracis-33583 384 10 # # SYM bracis-33583 384 11 2022/02981 2022/02981 NUM bracis-33583 384 12 - - SYM bracis-33583 384 13 8 8 NUM bracis-33583 384 14 . . PUNCT bracis-33583 385 1 author author NOUN bracis-33583 385 2 information information NOUN bracis-33583 385 3 authors author NOUN bracis-33583 385 4 and and CCONJ bracis-33583 385 5 affiliations affiliation NOUN bracis-33583 385 6 department department PROPN bracis-33583 385 7 of of ADP bracis-33583 385 8 computing computing PROPN bracis-33583 385 9 , , PUNCT bracis-33583 385 10 federal federal ADJ bracis-33583 385 11 university university PROPN bracis-33583 385 12 of of ADP bracis-33583 385 13 são são PROPN bracis-33583 385 14 carlos carlos PROPN bracis-33583 385 15 , , PUNCT bracis-33583 385 16 são são PROPN bracis-33583 385 17 carlos carlos PROPN bracis-33583 385 18 , , PUNCT bracis-33583 385 19 sp sp PROPN bracis-33583 385 20 , , PUNCT bracis-33583 385 21 brazil brazil PROPN bracis-33583 385 22 pedro pedro PROPN bracis-33583 385 23 henrique henrique PROPN bracis-33583 385 24 casarotto casarotto PROPN bracis-33583 385 25 institute institute PROPN bracis-33583 385 26 of of ADP bracis-33583 385 27 mathematics mathematics PROPN bracis-33583 385 28 and and CCONJ bracis-33583 385 29 computer computer NOUN bracis-33583 385 30 science science NOUN bracis-33583 385 31 , , PUNCT bracis-33583 385 32 university university PROPN bracis-33583 385 33 of of ADP bracis-33583 385 34 são são PROPN bracis-33583 385 35 paulo paulo PROPN bracis-33583 385 36 , , PUNCT bracis-33583 385 37 são são PROPN bracis-33583 385 38 carlos carlos PROPN bracis-33583 385 39 , , PUNCT bracis-33583 385 40 sp sp PROPN bracis-33583 385 41 , , PUNCT bracis-33583 385 42 brazil brazil PROPN bracis-33583 385 43 ricardo ricardo PROPN bracis-33583 385 44 cerri cerri PROPN bracis-33583 385 45 authors author NOUN bracis-33583 385 46 pedro pedro VERB bracis-33583 385 47 henrique henrique PROPN bracis-33583 385 48 casarottoview casarottoview PROPN bracis-33583 385 49 author author NOUN bracis-33583 385 50 publications publication NOUN bracis-33583 385 51 search search NOUN bracis-33583 385 52 author author NOUN bracis-33583 385 53 on on ADP bracis-33583 385 54 : : PUNCT bracis-33583 385 55 pubmed pubmed PROPN bracis-33583 385 56     SPACE bracis-33583 385 57 google google PROPN bracis-33583 385 58 scholar scholar NOUN bracis-33583 385 59 ricardo ricardo PROPN bracis-33583 385 60 cerriview cerriview NOUN bracis-33583 385 61 author author NOUN bracis-33583 385 62 publications publication NOUN bracis-33583 385 63 search search NOUN bracis-33583 385 64 author author NOUN bracis-33583 385 65 on on ADP bracis-33583 385 66 : : PUNCT bracis-33583 385 67 pubmed pubmed PROPN bracis-33583 385 68     SPACE bracis-33583 385 69 google google PROPN bracis-33583 385 70 scholar scholar NOUN bracis-33583 385 71 corresponding correspond VERB bracis-33583 385 72 author author NOUN bracis-33583 385 73 correspondence correspondence NOUN bracis-33583 385 74 to to ADP bracis-33583 385 75 ricardo ricardo PROPN bracis-33583 385 76 cerri cerri PROPN bracis-33583 385 77 . . PUNCT bracis-33583 386 1 editor editor NOUN bracis-33583 386 2 information information NOUN bracis-33583 386 3 editors editor NOUN bracis-33583 386 4 and and CCONJ bracis-33583 386 5 affiliations affiliation NOUN bracis-33583 386 6 universidade universidade VERB bracis-33583 386 7 federal federal PROPN bracis-33583 386 8 fluminense fluminense PROPN bracis-33583 386 9 , , PUNCT bracis-33583 386 10 niterói niterói PROPN bracis-33583 386 11 , , PUNCT bracis-33583 386 12 brazil brazil PROPN bracis-33583 386 13 aline aline PROPN bracis-33583 386 14 paes paes PROPN bracis-33583 386 15 instituto instituto PROPN bracis-33583 386 16 tecnológico tecnológico PROPN bracis-33583 386 17 de de PROPN bracis-33583 386 18 aeronáutica aeronáutica PROPN bracis-33583 386 19 , , PUNCT bracis-33583 386 20 são são PROPN bracis-33583 386 21 josé josé PROPN bracis-33583 386 22 dos dos PROPN bracis-33583 386 23 campos campos PROPN bracis-33583 386 24 , , PUNCT bracis-33583 386 25 brazil brazil PROPN bracis-33583 386 26 filipe filipe PROPN bracis-33583 386 27 a. a. PROPN bracis-33583 386 28 n. n. PROPN bracis-33583 386 29 verri verri PROPN bracis-33583 386 30 rights right NOUN bracis-33583 386 31 and and CCONJ bracis-33583 386 32 permissions permission NOUN bracis-33583 386 33 reprints reprint NOUN bracis-33583 386 34 and and CCONJ bracis-33583 386 35 permissions permission VERB bracis-33583 386 36 copyright copyright NOUN bracis-33583 386 37 information information NOUN bracis-33583 386 38 © © ADP bracis-33583 386 39 2025 2025 NUM bracis-33583 386 40 the the DET bracis-33583 386 41 author(s author(s NOUN bracis-33583 386 42 ) ) PUNCT bracis-33583 386 43 , , PUNCT bracis-33583 386 44 under under ADP bracis-33583 386 45 exclusive exclusive ADJ bracis-33583 386 46 license license NOUN bracis-33583 386 47 to to ADP bracis-33583 386 48 springer springer NOUN bracis-33583 386 49 nature nature PROPN bracis-33583 386 50 switzerland switzerland PROPN bracis-33583 386 51 ag ag PROPN bracis-33583 386 52 about about ADP bracis-33583 386 53 this this DET bracis-33583 386 54 paper paper NOUN bracis-33583 386 55 cite cite VERB bracis-33583 386 56 this this DET bracis-33583 386 57 paper paper NOUN bracis-33583 386 58 henrique henrique PROPN bracis-33583 386 59 casarotto casarotto PROPN bracis-33583 386 60 , , PUNCT bracis-33583 386 61 p. p. NOUN bracis-33583 386 62 , , PUNCT bracis-33583 386 63 cerri cerri PROPN bracis-33583 386 64 , , PUNCT bracis-33583 386 65 r. r. PROPN bracis-33583 386 66 ( ( PUNCT bracis-33583 386 67 2025 2025 NUM bracis-33583 386 68 ) ) PUNCT bracis-33583 386 69 . . PUNCT bracis-33583 387 1 growing grow VERB bracis-33583 387 2 self self NOUN bracis-33583 387 3 - - PUNCT bracis-33583 387 4 organizing organize VERB bracis-33583 387 5 maps map NOUN bracis-33583 387 6 for for ADP bracis-33583 387 7     SPACE bracis-33583 387 8 multi multi ADJ bracis-33583 387 9 - - ADJ bracis-33583 387 10 label label ADJ bracis-33583 387 11 classification classification NOUN bracis-33583 387 12 . . PUNCT bracis-33583 388 1 in in ADP bracis-33583 388 2 : : PUNCT bracis-33583 388 3 paes paes PROPN bracis-33583 388 4 , , PUNCT bracis-33583 388 5 a. a. PROPN bracis-33583 388 6 , , PUNCT bracis-33583 388 7 verri verri NOUN bracis-33583 388 8 , , PUNCT bracis-33583 388 9 f.a.n f.a.n ADJ bracis-33583 388 10 . . PUNCT bracis-33583 388 11 ( ( PUNCT bracis-33583 388 12 eds ed NOUN bracis-33583 388 13 ) ) PUNCT bracis-33583 388 14 intelligent intelligent ADJ bracis-33583 388 15 systems system NOUN bracis-33583 388 16 . . PUNCT bracis-33583 388 17 bracis bracis PROPN bracis-33583 388 18 2024 2024 NUM bracis-33583 388 19 . . PUNCT bracis-33583 389 1 lecture lecture NOUN bracis-33583 389 2 notes note NOUN bracis-33583 389 3 in in ADP bracis-33583 389 4 computer computer NOUN bracis-33583 389 5 science science NOUN bracis-33583 389 6 ( ( PUNCT bracis-33583 389 7 ) ) PUNCT bracis-33583 389 8 , , PUNCT bracis-33583 389 9 vol vol NOUN bracis-33583 389 10 15413 15413 NUM bracis-33583 389 11 . . PUNCT bracis-33583 390 1 springer springer NOUN bracis-33583 390 2 , , PUNCT bracis-33583 390 3 cham cham PROPN bracis-33583 390 4 . . PUNCT bracis-33583 391 1 https://doi.org/10.1007/978-3-031-79032-4_3 https://doi.org/10.1007/978-3-031-79032-4_3 PROPN bracis-33583 391 2 download download PROPN bracis-33583 391 3 citation citation NOUN bracis-33583 391 4 .ris .ris PUNCT bracis-33583 391 5 .enw .enw PROPN bracis-33583 391 6 .bib .bib PUNCT bracis-33583 392 1 doi doi PROPN bracis-33583 392 2 : : PUNCT bracis-33583 392 3 https://doi.org/10.1007/978-3-031-79032-4_3 https://doi.org/10.1007/978-3-031-79032-4_3 NUM bracis-33583 392 4 published publish VERB bracis-33583 392 5 : : PUNCT bracis-33583 392 6 30 30 NUM bracis-33583 392 7 january january PROPN bracis-33583 392 8 2025 2025 NUM bracis-33583 392 9 publisher publisher NOUN bracis-33583 392 10 name name NOUN bracis-33583 392 11 : : PUNCT bracis-33583 392 12 springer springer NOUN bracis-33583 392 13 , , PUNCT bracis-33583 392 14 cham cham PROPN bracis-33583 392 15 print print PROPN bracis-33583 392 16 isbn isbn PROPN bracis-33583 392 17 : : PUNCT bracis-33583 392 18 978 978 NUM bracis-33583 392 19 - - SYM bracis-33583 392 20 3 3 NUM bracis-33583 392 21 - - PUNCT bracis-33583 392 22 031 031 NUM bracis-33583 392 23 - - PUNCT bracis-33583 392 24 79031 79031 NUM bracis-33583 392 25 - - SYM bracis-33583 392 26 7 7 NUM bracis-33583 392 27 online online ADJ bracis-33583 392 28 isbn isbn NOUN bracis-33583 392 29 : : PUNCT bracis-33583 392 30 978 978 NUM bracis-33583 392 31 - - SYM bracis-33583 392 32 3 3 NUM bracis-33583 392 33 - - PUNCT bracis-33583 392 34 031 031 NUM bracis-33583 392 35 - - PUNCT bracis-33583 392 36 79032 79032 NUM bracis-33583 392 37 - - SYM bracis-33583 392 38 4 4 NUM bracis-33583 392 39 ebook ebook NOUN bracis-33583 392 40 packages package NOUN bracis-33583 392 41 : : PUNCT bracis-33583 392 42 computer computer NOUN bracis-33583 392 43 sciencecomputer sciencecomputer NOUN bracis-33583 392 44 science science NOUN bracis-33583 392 45 ( ( PUNCT bracis-33583 392 46 r0 r0 NOUN bracis-33583 392 47 ) ) PUNCT bracis-33583 392 48 share share VERB bracis-33583 392 49 this this DET bracis-33583 392 50 paper paper NOUN bracis-33583 392 51 anyone anyone PRON bracis-33583 392 52 you you PRON bracis-33583 392 53 share share VERB bracis-33583 392 54 the the DET bracis-33583 392 55 following follow VERB bracis-33583 392 56 link link NOUN bracis-33583 392 57 with with ADP bracis-33583 392 58 will will AUX bracis-33583 392 59 be be AUX bracis-33583 392 60 able able ADJ bracis-33583 392 61 to to PART bracis-33583 392 62 read read VERB bracis-33583 392 63 this this DET bracis-33583 392 64 content content NOUN bracis-33583 392 65 : : PUNCT bracis-33583 392 66 get get VERB bracis-33583 392 67 shareable shareable ADJ bracis-33583 392 68 linksorry linksorry NOUN bracis-33583 392 69 , , PUNCT bracis-33583 392 70 a a DET bracis-33583 392 71 shareable shareable ADJ bracis-33583 392 72 link link NOUN bracis-33583 392 73 is be AUX bracis-33583 392 74 not not PART bracis-33583 392 75 currently currently ADV bracis-33583 392 76 available available ADJ bracis-33583 392 77 for for ADP bracis-33583 392 78 this this DET bracis-33583 392 79 article article NOUN bracis-33583 392 80 . . PUNCT bracis-33583 393 1 copy copy VERB bracis-33583 393 2 shareable shareable ADJ bracis-33583 393 3 link link NOUN bracis-33583 393 4 to to PART bracis-33583 393 5 clipboard clipboard NOUN bracis-33583 393 6 provided provide VERB bracis-33583 393 7 by by ADP bracis-33583 393 8 the the DET bracis-33583 393 9 springer springer NOUN bracis-33583 393 10 nature nature PROPN bracis-33583 393 11 sharedit sharedit PROPN bracis-33583 393 12 content content NOUN bracis-33583 393 13 - - PUNCT bracis-33583 393 14 sharing share VERB bracis-33583 393 15 initiative initiative NOUN bracis-33583 393 16 keywords keyword NOUN bracis-33583 393 17 growing grow VERB bracis-33583 393 18 self self NOUN bracis-33583 393 19 - - PUNCT bracis-33583 393 20 organizing organize VERB bracis-33583 393 21 maps map NOUN bracis-33583 393 22 multi multi ADJ bracis-33583 393 23 - - ADJ bracis-33583 393 24 label label ADJ bracis-33583 393 25 classification classification NOUN bracis-33583 393 26 publish publish NOUN bracis-33583 393 27 with with ADP bracis-33583 393 28 us us PROPN bracis-33583 393 29 policies policy NOUN bracis-33583 393 30 and and CCONJ bracis-33583 393 31 ethics ethic NOUN bracis-33583 393 32 profiles profile NOUN bracis-33583 393 33 ricardo ricardo PROPN bracis-33583 393 34 cerri cerri PROPN bracis-33583 393 35 view view PROPN bracis-33583 393 36 author author NOUN bracis-33583 393 37 profile profile NOUN bracis-33583 393 38 search search NOUN bracis-33583 393 39 search search NOUN bracis-33583 393 40 by by ADP bracis-33583 393 41 keyword keyword NOUN bracis-33583 393 42 or or CCONJ bracis-33583 393 43 author author NOUN bracis-33583 393 44 search search NOUN bracis-33583 393 45 navigation navigation NOUN bracis-33583 393 46 find find VERB bracis-33583 393 47 a a DET bracis-33583 393 48 journal journal NOUN bracis-33583 393 49 publish publish VERB bracis-33583 393 50 with with ADP bracis-33583 393 51 us we PRON bracis-33583 393 52 track track VERB bracis-33583 393 53 your your PRON bracis-33583 393 54 research research NOUN bracis-33583 393 55 discover discover VERB bracis-33583 393 56 content content NOUN bracis-33583 393 57 journals journal NOUN bracis-33583 393 58 a a DET bracis-33583 393 59 - - PUNCT bracis-33583 393 60 z z NOUN bracis-33583 393 61 books book NOUN bracis-33583 393 62 a a DET bracis-33583 393 63 - - PUNCT bracis-33583 393 64 z z NOUN bracis-33583 393 65 publish publish NOUN bracis-33583 393 66 with with ADP bracis-33583 393 67 us us PROPN bracis-33583 393 68 journal journal PROPN bracis-33583 393 69 finder finder PROPN bracis-33583 393 70 publish publish VERB bracis-33583 393 71 your your PRON bracis-33583 393 72 research research NOUN bracis-33583 393 73 language language NOUN bracis-33583 393 74 editing edit VERB bracis-33583 393 75 open open ADJ bracis-33583 393 76 access access NOUN bracis-33583 393 77 publishing publishing NOUN bracis-33583 393 78 products product NOUN bracis-33583 393 79 and and CCONJ bracis-33583 393 80 services service NOUN bracis-33583 393 81 our our PRON bracis-33583 393 82 products product NOUN bracis-33583 393 83 librarians librarian VERB bracis-33583 393 84 societies society NOUN bracis-33583 393 85 partners partner NOUN bracis-33583 393 86 and and CCONJ bracis-33583 393 87 advertisers advertiser NOUN bracis-33583 393 88 our our PRON bracis-33583 393 89 brands brand NOUN bracis-33583 393 90 springer springer NOUN bracis-33583 393 91 nature nature PROPN bracis-33583 393 92 portfolio portfolio PROPN bracis-33583 393 93 bmc bmc PROPN bracis-33583 393 94 palgrave palgrave PROPN bracis-33583 393 95 macmillan macmillan PROPN bracis-33583 393 96 apress apress PROPN bracis-33583 393 97 discover discover VERB bracis-33583 393 98 your your PRON bracis-33583 393 99 privacy privacy NOUN bracis-33583 393 100 choices choice NOUN bracis-33583 393 101 / / SYM bracis-33583 393 102 manage manage NOUN bracis-33583 393 103 cookies cookie NOUN bracis-33583 393 104 your your PRON bracis-33583 393 105 us us PROPN bracis-33583 394 1 state state NOUN bracis-33583 394 2 privacy privacy NOUN bracis-33583 394 3 rights right NOUN bracis-33583 394 4 accessibility accessibility NOUN bracis-33583 394 5 statement statement NOUN bracis-33583 394 6 terms term NOUN bracis-33583 394 7 and and CCONJ bracis-33583 394 8 conditions condition NOUN bracis-33583 394 9 privacy privacy NOUN bracis-33583 394 10 policy policy NOUN bracis-33583 394 11 help help NOUN bracis-33583 394 12 and and CCONJ bracis-33583 394 13 support support VERB bracis-33583 394 14 legal legal ADJ bracis-33583 394 15 notice notice NOUN bracis-33583 394 16 cancel cancel VERB bracis-33583 394 17 contracts contract NOUN bracis-33583 394 18 here here ADV bracis-33583 394 19 129.74.145.123 129.74.145.123 NUM bracis-33583 394 20 hesburgh hesburgh PROPN bracis-33583 394 21 library library PROPN bracis-33583 394 22 er er INTJ bracis-33583 394 23 unit unit NOUN bracis-33583 394 24 ( ( PUNCT bracis-33583 394 25 3005732405 3005732405 NUM bracis-33583 394 26 ) ) PUNCT bracis-33583 394 27 northeast northeast ADJ bracis-33583 394 28 research research NOUN bracis-33583 394 29 libraries library NOUN bracis-33583 394 30 ( ( PUNCT bracis-33583 394 31 nerl nerl PROPN bracis-33583 394 32 ) ) PUNCT bracis-33583 394 33 ( ( PUNCT bracis-33583 394 34 8200828607 8200828607 NUM bracis-33583 394 35 ) ) PUNCT bracis-33583 394 36 nerl nerl VERB bracis-33583 394 37 ta ta X bracis-33583 394 38 account account NOUN bracis-33583 394 39 ( ( PUNCT bracis-33583 394 40 3006206169 3006206169 NUM bracis-33583 394 41 ) ) PUNCT bracis-33583 394 42 university university NOUN bracis-33583 394 43 of of ADP bracis-33583 394 44 notre notre PROPN bracis-33583 394 45 dame dame PROPN bracis-33583 394 46 hesburgh hesburgh PROPN bracis-33583 394 47 library library NOUN bracis-33583 394 48 ( ( PUNCT bracis-33583 394 49 3000184373 3000184373 NUM bracis-33583 394 50 ) ) PUNCT bracis-33583 395 1 © © ADP bracis-33583 395 2 2025 2025 NUM bracis-33583 395 3 springer springer NOUN bracis-33583 395 4 nature nature NOUN