id sid tid token lemma pos iajs-3467 1 1 412 412 NUM iajs-3467 1 2 ihjpas ihjpa NOUN iajs-3467 1 3 . . PUNCT iajs-3467 2 1 37 37 NUM iajs-3467 2 2 ( ( PUNCT iajs-3467 2 3 1 1 NUM iajs-3467 2 4 ) ) PUNCT iajs-3467 2 5 2024 2024 NUM iajs-3467 2 6 ibn ibn PROPN iajs-3467 2 7 al al PROPN iajs-3467 2 8 - - PUNCT iajs-3467 2 9 haitham haitham PROPN iajs-3467 2 10 journal journal PROPN iajs-3467 2 11 for for ADP iajs-3467 2 12 pure pure ADJ iajs-3467 2 13 and and CCONJ iajs-3467 2 14 applied applied ADJ iajs-3467 2 15 sciences sciences PROPN iajs-3467 2 16 journal journal PROPN iajs-3467 2 17 homepage homepage NOUN iajs-3467 2 18 : : PUNCT iajs-3467 2 19 jih.uobaghdad.edu.iq jih.uobaghdad.edu.iq NOUN iajs-3467 2 20 pissn pissn ADJ iajs-3467 2 21 : : PUNCT iajs-3467 2 22 1609 1609 NUM iajs-3467 2 23 - - SYM iajs-3467 2 24 4042 4042 NUM iajs-3467 2 25 , , PUNCT iajs-3467 2 26 eissn eissn NOUN iajs-3467 2 27 : : PUNCT iajs-3467 2 28 2521 2521 NUM iajs-3467 2 29 - - SYM iajs-3467 2 30 3407 3407 NUM iajs-3467 2 31 1seror 1seror NUM iajs-3467 2 32 faeq faeq NOUN iajs-3467 2 33 mohammed mohammed PROPN iajs-3467 2 34 2ghadeer 2ghadeer PROPN iajs-3467 2 35 jasim jasim PROPN iajs-3467 2 36 mohammed mohammed PROPN iajs-3467 2 37 mahdi mahdi PROPN iajs-3467 2 38 * * PROPN iajs-3467 2 39 3md 3md PROPN iajs-3467 2 40 kamrul kamrul PROPN iajs-3467 2 41 hasan hasan PROPN iajs-3467 2 42 khan khan PROPN iajs-3467 3 1 1,2department 1,2department NUM iajs-3467 3 2 of of ADP iajs-3467 3 3 mathematics mathematic NOUN iajs-3467 3 4 , , PUNCT iajs-3467 3 5 college college NOUN iajs-3467 3 6 of of ADP iajs-3467 3 7 education education NOUN iajs-3467 3 8 for for ADP iajs-3467 3 9 the the DET iajs-3467 3 10 pure pure ADJ iajs-3467 3 11 science science NOUN iajs-3467 3 12 ibn ibn PROPN iajs-3467 3 13 al al PROPN iajs-3467 3 14 - - PUNCT iajs-3467 3 15 haitham haitham PROPN iajs-3467 3 16 , , PUNCT iajs-3467 3 17 university university PROPN iajs-3467 3 18 of of ADP iajs-3467 3 19 baghdad baghdad PROPN iajs-3467 3 20 , , PUNCT iajs-3467 3 21 iraq iraq PROPN iajs-3467 3 22 , , PUNCT iajs-3467 3 23 . . PUNCT iajs-3467 4 1 3department 3department NUM iajs-3467 4 2 of of ADP iajs-3467 4 3 mathematical mathematical ADJ iajs-3467 4 4 sciences sciences PROPN iajs-3467 4 5 , , PUNCT iajs-3467 4 6 fulbright fulbright PROPN iajs-3467 4 7 college college PROPN iajs-3467 4 8 of of ADP iajs-3467 4 9 arts arts PROPN iajs-3467 4 10 & & CCONJ iajs-3467 4 11 sciences sciences PROPN iajs-3467 4 12 , , PUNCT iajs-3467 4 13 1university 1university NUM iajs-3467 4 14 of of ADP iajs-3467 4 15 arkansas arkansas PROPN iajs-3467 4 16 , , PUNCT iajs-3467 4 17 usa usa PROPN iajs-3467 4 18 * * PUNCT iajs-3467 4 19 corresponding correspond VERB iajs-3467 4 20 author author NOUN iajs-3467 4 21 . . PUNCT iajs-3467 5 1 gmahdi@ihcoedu.uobaghdad.edu.iq gmahdi@ihcoedu.uobaghdad.edu.iq NOUN iajs-3467 5 2 , , PUNCT iajs-3467 5 3 abstract abstract ADJ iajs-3467 5 4 support support NOUN iajs-3467 5 5 vector vector NOUN iajs-3467 5 6 machines machine NOUN iajs-3467 5 7 ( ( PUNCT iajs-3467 5 8 svms svms NOUN iajs-3467 5 9 ) ) PUNCT iajs-3467 5 10 are be AUX iajs-3467 5 11 supervised supervise VERB iajs-3467 5 12 learning learning NOUN iajs-3467 5 13 models model NOUN iajs-3467 5 14 used use VERB iajs-3467 5 15 to to PART iajs-3467 5 16 examine examine VERB iajs-3467 5 17 data datum NOUN iajs-3467 5 18 sets set NOUN iajs-3467 5 19 in in ADP iajs-3467 5 20 order order NOUN iajs-3467 5 21 to to PART iajs-3467 5 22 classify classify VERB iajs-3467 5 23 or or CCONJ iajs-3467 5 24 predict predict VERB iajs-3467 5 25 dependent dependent ADJ iajs-3467 5 26 variables variable NOUN iajs-3467 5 27 . . PUNCT iajs-3467 6 1 svm svm PROPN iajs-3467 6 2 is be AUX iajs-3467 6 3 typically typically ADV iajs-3467 6 4 used use VERB iajs-3467 6 5 for for ADP iajs-3467 6 6 classification classification NOUN iajs-3467 6 7 by by ADP iajs-3467 6 8 determining determine VERB iajs-3467 6 9 the the DET iajs-3467 6 10 best good ADJ iajs-3467 6 11 hyperplane hyperplane NOUN iajs-3467 6 12 between between ADP iajs-3467 6 13 two two NUM iajs-3467 6 14 classes class NOUN iajs-3467 6 15 . . PUNCT iajs-3467 7 1 however however ADV iajs-3467 7 2 , , PUNCT iajs-3467 7 3 working work VERB iajs-3467 7 4 with with ADP iajs-3467 7 5 huge huge ADJ iajs-3467 7 6 datasets dataset NOUN iajs-3467 7 7 can can AUX iajs-3467 7 8 lead lead VERB iajs-3467 7 9 to to ADP iajs-3467 7 10 a a DET iajs-3467 7 11 number number NOUN iajs-3467 7 12 of of ADP iajs-3467 7 13 problems problem NOUN iajs-3467 7 14 , , PUNCT iajs-3467 7 15 including include VERB iajs-3467 7 16 time time NOUN iajs-3467 7 17 - - PUNCT iajs-3467 7 18 consuming consume VERB iajs-3467 7 19 and and CCONJ iajs-3467 7 20 inefficient inefficient ADJ iajs-3467 7 21 solutions solution NOUN iajs-3467 7 22 . . PUNCT iajs-3467 8 1 this this DET iajs-3467 8 2 research research NOUN iajs-3467 8 3 updates update VERB iajs-3467 8 4 the the DET iajs-3467 8 5 svm svm NOUN iajs-3467 8 6 by by ADP iajs-3467 8 7 employing employ VERB iajs-3467 8 8 a a DET iajs-3467 8 9 stochastic stochastic ADJ iajs-3467 8 10 gradient gradient ADJ iajs-3467 8 11 descent descent NOUN iajs-3467 8 12 method method NOUN iajs-3467 8 13 . . PUNCT iajs-3467 9 1 the the DET iajs-3467 9 2 new new ADJ iajs-3467 9 3 approach approach NOUN iajs-3467 9 4 , , PUNCT iajs-3467 9 5 the the DET iajs-3467 9 6 extended extended ADJ iajs-3467 9 7 stochastic stochastic ADJ iajs-3467 9 8 gradient gradient ADJ iajs-3467 9 9 descent descent NOUN iajs-3467 9 10 svm svm NOUN iajs-3467 9 11 ( ( PUNCT iajs-3467 9 12 esgd esgd NOUN iajs-3467 9 13 - - PUNCT iajs-3467 9 14 svm svm NOUN iajs-3467 9 15 ) ) PUNCT iajs-3467 9 16 , , PUNCT iajs-3467 9 17 was be AUX iajs-3467 9 18 tested test VERB iajs-3467 9 19 on on ADP iajs-3467 9 20 two two NUM iajs-3467 9 21 simulation simulation NOUN iajs-3467 9 22 datasets dataset NOUN iajs-3467 9 23 . . PUNCT iajs-3467 10 1 the the DET iajs-3467 10 2 proposed propose VERB iajs-3467 10 3 method method NOUN iajs-3467 10 4 was be AUX iajs-3467 10 5 compared compare VERB iajs-3467 10 6 with with ADP iajs-3467 10 7 other other ADJ iajs-3467 10 8 classification classification NOUN iajs-3467 10 9 approaches approach NOUN iajs-3467 10 10 such such ADJ iajs-3467 10 11 as as ADP iajs-3467 10 12 logistic logistic ADJ iajs-3467 10 13 regression regression NOUN iajs-3467 10 14 , , PUNCT iajs-3467 10 15 naive naive ADJ iajs-3467 10 16 model model NOUN iajs-3467 10 17 , , PUNCT iajs-3467 10 18 k k PROPN iajs-3467 10 19 nearest near ADJ iajs-3467 10 20 neighbors neighbor NOUN iajs-3467 10 21 and and CCONJ iajs-3467 10 22 random random ADJ iajs-3467 10 23 forest forest NOUN iajs-3467 10 24 . . PUNCT iajs-3467 11 1 the the DET iajs-3467 11 2 results result NOUN iajs-3467 11 3 show show VERB iajs-3467 11 4 that that SCONJ iajs-3467 11 5 the the DET iajs-3467 11 6 esgd esgd NOUN iajs-3467 11 7 - - PUNCT iajs-3467 11 8 svm svm PROPN iajs-3467 11 9 has have VERB iajs-3467 11 10 a a DET iajs-3467 11 11 very very ADV iajs-3467 11 12 high high ADJ iajs-3467 11 13 accuracy accuracy NOUN iajs-3467 11 14 and and CCONJ iajs-3467 11 15 is be AUX iajs-3467 11 16 quite quite ADV iajs-3467 11 17 robust robust ADJ iajs-3467 11 18 . . PUNCT iajs-3467 12 1 esgd esgd NOUN iajs-3467 12 2 - - PUNCT iajs-3467 12 3 svm svm PROPN iajs-3467 12 4 is be AUX iajs-3467 12 5 used use VERB iajs-3467 12 6 to to PART iajs-3467 12 7 analyze analyze VERB iajs-3467 12 8 the the DET iajs-3467 12 9 heart heart NOUN iajs-3467 12 10 disease disease NOUN iajs-3467 12 11 dataset dataset VERB iajs-3467 12 12 downloaded download VERB iajs-3467 12 13 from from ADP iajs-3467 12 14 harvard harvard PROPN iajs-3467 12 15 dataverse dataverse PROPN iajs-3467 12 16 . . PUNCT iajs-3467 13 1 the the DET iajs-3467 13 2 entire entire ADJ iajs-3467 13 3 analysis analysis NOUN iajs-3467 13 4 was be AUX iajs-3467 13 5 performed perform VERB iajs-3467 13 6 using use VERB iajs-3467 13 7 the the DET iajs-3467 13 8 program program NOUN iajs-3467 13 9 r r NOUN iajs-3467 13 10 version version NOUN iajs-3467 13 11 4.3 4.3 NUM iajs-3467 13 12 . . PUNCT iajs-3467 14 1 keywords keyword NOUN iajs-3467 14 2 svm svm VERB iajs-3467 14 3 , , PUNCT iajs-3467 14 4 classification classification NOUN iajs-3467 14 5 , , PUNCT iajs-3467 14 6 reduction reduction NOUN iajs-3467 14 7 of of ADP iajs-3467 14 8 dimensions dimension NOUN iajs-3467 14 9 , , PUNCT iajs-3467 14 10 variables variable NOUN iajs-3467 14 11 selection selection NOUN iajs-3467 14 12 , , PUNCT iajs-3467 14 13 gradient gradient ADJ iajs-3467 14 14 descent descent NOUN iajs-3467 14 15 , , PUNCT iajs-3467 14 16 heart heart NOUN iajs-3467 14 17 disease disease NOUN iajs-3467 14 18 . . PUNCT iajs-3467 15 1 1 1 X iajs-3467 15 2 . . X iajs-3467 15 3 introduction introduction NOUN iajs-3467 15 4 suppose suppose VERB iajs-3467 15 5 we we PRON iajs-3467 15 6 have have VERB iajs-3467 15 7 a a DET iajs-3467 15 8 data data NOUN iajs-3467 15 9 set set VERB iajs-3467 15 10 where where SCONJ iajs-3467 15 11 for for ADP iajs-3467 15 12 each each DET iajs-3467 15 13 subject subject NOUN iajs-3467 15 14 we we PRON iajs-3467 15 15 have have VERB iajs-3467 15 16 information information NOUN iajs-3467 15 17 about about ADP iajs-3467 15 18 an an DET iajs-3467 15 19 𝑝dimensional 𝑝dimensional ADJ iajs-3467 15 20 covariate covariate ADJ iajs-3467 15 21 vector vector NOUN iajs-3467 15 22 , , PUNCT iajs-3467 15 23 π‘₯𝑝×1 π‘₯𝑝×1 PROPN iajs-3467 15 24 and and CCONJ iajs-3467 15 25 a a DET iajs-3467 15 26 response response NOUN iajs-3467 15 27 y y PROPN iajs-3467 15 28 that that PRON iajs-3467 15 29 has have VERB iajs-3467 15 30 two two NUM iajs-3467 15 31 possible possible ADJ iajs-3467 15 32 categories category NOUN iajs-3467 15 33 . . PUNCT iajs-3467 16 1 our our PRON iajs-3467 16 2 goal goal NOUN iajs-3467 16 3 is be AUX iajs-3467 16 4 to to PART iajs-3467 16 5 develop develop VERB iajs-3467 16 6 an an DET iajs-3467 16 7 algorithm algorithm NOUN iajs-3467 16 8 that that PRON iajs-3467 16 9 allows allow VERB iajs-3467 16 10 us we PRON iajs-3467 16 11 to to PART iajs-3467 16 12 predict predict VERB iajs-3467 16 13 for for ADP iajs-3467 16 14 each each DET iajs-3467 16 15 new new ADJ iajs-3467 16 16 observation observation NOUN iajs-3467 16 17 the the DET iajs-3467 16 18 category category NOUN iajs-3467 16 19 of of ADP iajs-3467 16 20 its its PRON iajs-3467 16 21 response response NOUN iajs-3467 16 22 based base VERB iajs-3467 16 23 on on ADP iajs-3467 16 24 its its PRON iajs-3467 16 25 𝑝-information 𝑝-information NOUN iajs-3467 16 26 . . PUNCT iajs-3467 17 1 the the DET iajs-3467 17 2 support support NOUN iajs-3467 17 3 vector vector NOUN iajs-3467 17 4 machine machine NOUN iajs-3467 17 5 ( ( PUNCT iajs-3467 17 6 svm svm PROPN iajs-3467 17 7 ) ) PUNCT iajs-3467 17 8 is be AUX iajs-3467 17 9 one one NUM iajs-3467 17 10 of of ADP iajs-3467 17 11 the the DET iajs-3467 17 12 many many ADJ iajs-3467 17 13 methods method NOUN iajs-3467 17 14 to to PART iajs-3467 17 15 do do AUX iajs-3467 17 16 this this PRON iajs-3467 17 17 , , PUNCT iajs-3467 17 18 such such ADJ iajs-3467 17 19 as as ADP iajs-3467 17 20 : : PUNCT iajs-3467 17 21 logistic logistic ADJ iajs-3467 17 22 regression regression NOUN iajs-3467 17 23 [ [ X iajs-3467 17 24 1 1 NUM iajs-3467 17 25 ] ] PUNCT iajs-3467 17 26 , , PUNCT iajs-3467 17 27 random random ADJ iajs-3467 17 28 forest forest NOUN iajs-3467 18 1 [ [ X iajs-3467 18 2 2 2 NUM iajs-3467 18 3 ] ] PUNCT iajs-3467 18 4 , , PUNCT iajs-3467 18 5 k k PROPN iajs-3467 18 6 nearest near ADJ iajs-3467 18 7 neighbors neighbor NOUN iajs-3467 18 8 [ [ X iajs-3467 18 9 3 3 NUM iajs-3467 18 10 ] ] PUNCT iajs-3467 18 11 , , PUNCT iajs-3467 18 12 naΓ―ve naΓ―ve ADJ iajs-3467 18 13 bayes baye NOUN iajs-3467 19 1 [ [ X iajs-3467 19 2 4 4 X iajs-3467 19 3 ] ] PUNCT iajs-3467 19 4 and and CCONJ iajs-3467 19 5 lda lda X iajs-3467 20 1 [ [ X iajs-3467 20 2 5 5 NUM iajs-3467 20 3 ] ] PUNCT iajs-3467 20 4 . . PUNCT iajs-3467 21 1 svm svm PROPN iajs-3467 21 2 is be AUX iajs-3467 21 3 a a DET iajs-3467 21 4 supervised supervised ADJ iajs-3467 21 5 learning learning NOUN iajs-3467 21 6 technique technique NOUN iajs-3467 21 7 , , PUNCT iajs-3467 21 8 i.e. i.e. X iajs-3467 21 9 , , PUNCT iajs-3467 21 10 we we PRON iajs-3467 21 11 create create VERB iajs-3467 21 12 a a DET iajs-3467 21 13 classifier classifier NOUN iajs-3467 21 14 based base VERB iajs-3467 21 15 on on ADP iajs-3467 21 16 a a DET iajs-3467 21 17 training training NOUN iajs-3467 21 18 dataset dataset NOUN iajs-3467 21 19 and and CCONJ iajs-3467 21 20 use use VERB iajs-3467 21 21 this this DET iajs-3467 21 22 classifier classifier NOUN iajs-3467 21 23 for for ADP iajs-3467 21 24 future future ADJ iajs-3467 21 25 observations observation NOUN iajs-3467 21 26 . . PUNCT iajs-3467 22 1 received receive VERB iajs-3467 22 2 7 7 NUM iajs-3467 22 3 may may PROPN iajs-3467 22 4 2023 2023 NUM iajs-3467 22 5 , , PUNCT iajs-3467 22 6 received receive VERB iajs-3467 22 7 8 8 NUM iajs-3467 22 8 june june PROPN iajs-3467 22 9 2023 2023 NUM iajs-3467 22 10 , , PUNCT iajs-3467 22 11 accepted accept VERB iajs-3467 22 12 12 12 NUM iajs-3467 22 13 june june PROPN iajs-3467 22 14 2023 2023 NUM iajs-3467 22 15 , , PUNCT iajs-3467 22 16 published publish VERB iajs-3467 22 17 20 20 NUM iajs-3467 22 18 january january PROPN iajs-3467 22 19 2024 2024 NUM iajs-3467 22 20 enhanced enhance VERB iajs-3467 22 21 support support NOUN iajs-3467 22 22 vector vector NOUN iajs-3467 22 23 machine machine NOUN iajs-3467 22 24 methods method NOUN iajs-3467 22 25 using use VERB iajs-3467 22 26 stochastic stochastic ADJ iajs-3467 22 27 gradient gradient ADJ iajs-3467 22 28 descent descent NOUN iajs-3467 22 29 and and CCONJ iajs-3467 22 30 its its PRON iajs-3467 22 31 application application NOUN iajs-3467 22 32 to to ADP iajs-3467 22 33 heart heart NOUN iajs-3467 22 34 disease disease NOUN iajs-3467 22 35 dataset dataset VERB iajs-3467 22 36 doi.org/10.30526/37.1.3467 doi.org/10.30526/37.1.3467 ADJ iajs-3467 22 37 https://jih.uobaghdad.edu.iq/index.php/j/index#1609-4042 https://jih.uobaghdad.edu.iq/index.php/j/index#1609-4042 NOUN iajs-3467 22 38 https://jih.uobaghdad.edu.iq/index.php/j/index#2521-3407 https://jih.uobaghdad.edu.iq/index.php/j/index#2521-3407 NOUN iajs-3467 22 39 mailto:gmahdi@ihcoedu.uobaghdad.edu.iq mailto:gmahdi@ihcoedu.uobaghdad.edu.iq NOUN iajs-3467 22 40 https://orcid.org/0000-0002-4444-4362 https://orcid.org/0000-0002-4444-4362 PROPN iajs-3467 22 41 mailto:serorfaeqmohammed@gmail.com mailto:serorfaeqmohammed@gmail.com PROPN iajs-3467 22 42 https://orcid.org/0000-0003-4870-4034 https://orcid.org/0000-0003-4870-4034 PROPN iajs-3467 22 43 mailto:gmahdi@ihcoedu.uobaghdad.edu.iq mailto:gmahdi@ihcoedu.uobaghdad.edu.iq PROPN iajs-3467 22 44 , , PUNCT iajs-3467 22 45 https://orcid.org/ https://orcid.org/ PRON iajs-3467 22 46 mailto:mkk006@uark.edu mailto:mkk006@uark.edu PROPN iajs-3467 22 47 ihjpas ihjpas PROPN iajs-3467 22 48 . . PUNCT iajs-3467 23 1 37 37 NUM iajs-3467 23 2 ( ( PUNCT iajs-3467 23 3 1 1 NUM iajs-3467 23 4 ) ) PUNCT iajs-3467 23 5 2024 2024 NUM iajs-3467 23 6 413 413 NUM iajs-3467 23 7 for for ADP iajs-3467 23 8 group group NOUN iajs-3467 23 9 one one NUM iajs-3467 23 10 , , PUNCT iajs-3467 23 11 let let VERB iajs-3467 23 12 𝑦 𝑦 NOUN iajs-3467 23 13 = = SYM iajs-3467 23 14 +1 +1 PROPN iajs-3467 23 15 and and CCONJ iajs-3467 23 16 for for ADP iajs-3467 23 17 group group NOUN iajs-3467 23 18 two two NUM iajs-3467 23 19 , , PUNCT iajs-3467 23 20 let let VERB iajs-3467 23 21 𝑦 𝑦 NOUN iajs-3467 23 22 = = NOUN iajs-3467 23 23 βˆ’1 βˆ’1 NOUN iajs-3467 23 24 . . PUNCT iajs-3467 24 1 the the DET iajs-3467 24 2 goal goal NOUN iajs-3467 24 3 of of ADP iajs-3467 24 4 svm svm PROPN iajs-3467 24 5 is be AUX iajs-3467 24 6 to to PART iajs-3467 24 7 design design VERB iajs-3467 24 8 a a DET iajs-3467 24 9 classifier classifier NOUN iajs-3467 24 10 𝑓(π‘₯ 𝑓(π‘₯ NOUN iajs-3467 24 11 ) ) PUNCT iajs-3467 24 12 such such ADJ iajs-3467 24 13 that that SCONJ iajs-3467 24 14 , , PUNCT iajs-3467 24 15 the the DET iajs-3467 24 16 classifier classifier NOUN iajs-3467 24 17 rule rule NOUN iajs-3467 24 18 : : PUNCT iajs-3467 24 19 𝑦 𝑦 NOUN iajs-3467 24 20 = = X iajs-3467 24 21 +1 +1 NOUN iajs-3467 24 22 if if SCONJ iajs-3467 24 23 𝑓(π‘₯ 𝑓(π‘₯ NOUN iajs-3467 24 24 ) ) PUNCT iajs-3467 24 25 > > X iajs-3467 24 26 0 0 PUNCT iajs-3467 24 27 and and CCONJ iajs-3467 24 28 𝑦 𝑦 NOUN iajs-3467 24 29 = = X iajs-3467 24 30 βˆ’1 βˆ’1 NOUN iajs-3467 24 31 if if SCONJ iajs-3467 24 32 𝑓(π‘₯ 𝑓(π‘₯ NOUN iajs-3467 24 33 ) ) PUNCT iajs-3467 24 34 < < X iajs-3467 24 35 0 0 NUM iajs-3467 24 36 , , PUNCT iajs-3467 24 37 can can AUX iajs-3467 24 38 be be AUX iajs-3467 24 39 used use VERB iajs-3467 24 40 to to PART iajs-3467 24 41 determine determine VERB iajs-3467 24 42 the the DET iajs-3467 24 43 response response NOUN iajs-3467 24 44 category category NOUN iajs-3467 24 45 given give VERB iajs-3467 24 46 the the DET iajs-3467 24 47 covariate covariate ADJ iajs-3467 24 48 information information NOUN iajs-3467 24 49 . . PUNCT iajs-3467 25 1 unlike unlike ADP iajs-3467 25 2 lda lda PROPN iajs-3467 25 3 , , PUNCT iajs-3467 25 4 svm svm PROPN iajs-3467 25 5 does do AUX iajs-3467 25 6 not not PART iajs-3467 25 7 use use VERB iajs-3467 25 8 any any DET iajs-3467 25 9 distribution distribution NOUN iajs-3467 25 10 for for ADP iajs-3467 25 11 π‘₯ π‘₯ PRON iajs-3467 25 12 given give VERB iajs-3467 25 13 its its PRON iajs-3467 25 14 category category NOUN iajs-3467 25 15 . . PUNCT iajs-3467 26 1 instead instead ADV iajs-3467 26 2 , , PUNCT iajs-3467 26 3 it it PRON iajs-3467 26 4 is be AUX iajs-3467 26 5 a a DET iajs-3467 26 6 geometric geometric ADJ iajs-3467 26 7 procedure procedure NOUN iajs-3467 26 8 that that PRON iajs-3467 26 9 finds find VERB iajs-3467 26 10 the the DET iajs-3467 26 11 classifier classifier NOUN iajs-3467 26 12 according accord VERB iajs-3467 26 13 to to ADP iajs-3467 26 14 some some DET iajs-3467 26 15 optimization optimization NOUN iajs-3467 26 16 criterion criterion NOUN iajs-3467 26 17 . . PUNCT iajs-3467 27 1 in in ADP iajs-3467 27 2 this this DET iajs-3467 27 3 work work NOUN iajs-3467 27 4 , , PUNCT iajs-3467 27 5 we we PRON iajs-3467 27 6 will will AUX iajs-3467 27 7 discuss discuss VERB iajs-3467 27 8 linear linear ADJ iajs-3467 27 9 and and CCONJ iajs-3467 27 10 nonlinear nonlinear ADJ iajs-3467 27 11 svmssuppose svmssuppose ADJ iajs-3467 27 12 𝑓(π‘₯)=,𝛽-0.+,𝛽-𝑇.𝑋 𝑓(π‘₯)=,𝛽-0.+,𝛽-𝑇.𝑋 NOUN iajs-3467 27 13 for for ADP iajs-3467 27 14 unknown unknown ADJ iajs-3467 27 15 parameters parameter NOUN iajs-3467 27 16 ( ( PUNCT iajs-3467 27 17 , , PUNCT iajs-3467 27 18 𝛽-0.,𝛽 𝛽-0.,𝛽 PROPN iajs-3467 27 19 ) ) PUNCT iajs-3467 27 20 is be AUX iajs-3467 27 21 a a DET iajs-3467 27 22 function function NOUN iajs-3467 27 23 , , PUNCT iajs-3467 27 24 a a DET iajs-3467 27 25 hyperplane hyperplane NOUN iajs-3467 27 26 in in ADP iajs-3467 27 27 the the DET iajs-3467 27 28 space space NOUN iajs-3467 27 29 , , PUNCT iajs-3467 27 30 that that PRON iajs-3467 27 31 acts act VERB iajs-3467 27 32 as as ADP iajs-3467 27 33 a a DET iajs-3467 27 34 separator separator NOUN iajs-3467 27 35 between between ADP iajs-3467 27 36 the the DET iajs-3467 27 37 two two NUM iajs-3467 27 38 response response NOUN iajs-3467 27 39 categories category NOUN iajs-3467 27 40 . . PUNCT iajs-3467 27 41 . . PUNCT iajs-3467 28 1 linearity linearity NOUN iajs-3467 28 2 is be AUX iajs-3467 28 3 a a DET iajs-3467 28 4 simplifying simplify VERB iajs-3467 28 5 assumption assumption NOUN iajs-3467 28 6 , , PUNCT iajs-3467 28 7 and and CCONJ iajs-3467 28 8 it it PRON iajs-3467 28 9 is be AUX iajs-3467 28 10 not not PART iajs-3467 28 11 reasonable reasonable ADJ iajs-3467 28 12 to to PART iajs-3467 28 13 always always ADV iajs-3467 28 14 assume assume VERB iajs-3467 28 15 that that SCONJ iajs-3467 28 16 Ζ’ Ζ’ PRON iajs-3467 28 17 is be AUX iajs-3467 28 18 linear linear ADJ iajs-3467 28 19 , , PUNCT iajs-3467 28 20 because because SCONJ iajs-3467 28 21 in in ADP iajs-3467 28 22 general general ADJ iajs-3467 28 23 , , PUNCT iajs-3467 28 24 non non ADJ iajs-3467 28 25 - - ADJ iajs-3467 28 26 linear linear ADJ iajs-3467 28 27 svms svms NOUN iajs-3467 28 28 are be AUX iajs-3467 28 29 a a DET iajs-3467 28 30 popular popular ADJ iajs-3467 28 31 tool tool NOUN iajs-3467 28 32 for for ADP iajs-3467 28 33 many many ADJ iajs-3467 28 34 real real ADJ iajs-3467 28 35 - - PUNCT iajs-3467 28 36 world world NOUN iajs-3467 28 37 applications application NOUN iajs-3467 28 38 . . PUNCT iajs-3467 29 1 in in ADP iajs-3467 29 2 some some DET iajs-3467 29 3 cases case NOUN iajs-3467 29 4 , , PUNCT iajs-3467 29 5 linear linear ADJ iajs-3467 29 6 separation separation NOUN iajs-3467 29 7 may may AUX iajs-3467 29 8 work work VERB iajs-3467 29 9 sufficiently sufficiently ADV iajs-3467 29 10 so so SCONJ iajs-3467 29 11 that that SCONJ iajs-3467 29 12 we we PRON iajs-3467 29 13 do do AUX iajs-3467 29 14 not not PART iajs-3467 29 15 need need VERB iajs-3467 29 16 to to PART iajs-3467 29 17 consider consider VERB iajs-3467 29 18 nonlinear nonlinear ADJ iajs-3467 29 19 assumptions assumption NOUN iajs-3467 29 20 . . PUNCT iajs-3467 30 1 we we PRON iajs-3467 30 2 organize organize VERB iajs-3467 30 3 the the DET iajs-3467 30 4 paper paper NOUN iajs-3467 30 5 as as SCONJ iajs-3467 30 6 follows follow VERB iajs-3467 30 7 : : PUNCT iajs-3467 30 8 in in ADP iajs-3467 30 9 sections section NOUN iajs-3467 30 10 2 2 NUM iajs-3467 30 11 and and CCONJ iajs-3467 30 12 3 3 NUM iajs-3467 30 13 , , PUNCT iajs-3467 30 14 we we PRON iajs-3467 30 15 introduce introduce VERB iajs-3467 30 16 the the DET iajs-3467 30 17 hard hard ADJ iajs-3467 30 18 ( ( PUNCT iajs-3467 30 19 linear linear ADJ iajs-3467 30 20 ) ) PUNCT iajs-3467 30 21 boundary boundary NOUN iajs-3467 30 22 and and CCONJ iajs-3467 30 23 the the DET iajs-3467 30 24 soft soft ADJ iajs-3467 30 25 ( ( PUNCT iajs-3467 30 26 nonlinear nonlinear ADJ iajs-3467 30 27 ) ) PUNCT iajs-3467 30 28 boundary boundary NOUN iajs-3467 30 29 . . PUNCT iajs-3467 31 1 the the DET iajs-3467 31 2 kernel kernel PROPN iajs-3467 31 3 transformation transformation NOUN iajs-3467 31 4 and and CCONJ iajs-3467 31 5 its its PRON iajs-3467 31 6 properties property NOUN iajs-3467 31 7 are be AUX iajs-3467 31 8 discussed discuss VERB iajs-3467 31 9 in in ADP iajs-3467 31 10 section section NOUN iajs-3467 31 11 4 4 NUM iajs-3467 31 12 . . PUNCT iajs-3467 32 1 the the DET iajs-3467 32 2 improved improve VERB iajs-3467 32 3 svm svm NOUN iajs-3467 32 4 with with ADP iajs-3467 32 5 stochastic stochastic ADJ iajs-3467 32 6 gradient gradient ADJ iajs-3467 32 7 descent descent NOUN iajs-3467 32 8 is be AUX iajs-3467 32 9 explained explain VERB iajs-3467 32 10 in in ADP iajs-3467 32 11 section section NOUN iajs-3467 32 12 5 5 NUM iajs-3467 32 13 . . PUNCT iajs-3467 32 14 simulation simulation NOUN iajs-3467 32 15 studies study NOUN iajs-3467 32 16 with with ADP iajs-3467 32 17 three three NUM iajs-3467 32 18 data datum NOUN iajs-3467 32 19 sets set NOUN iajs-3467 32 20 and and CCONJ iajs-3467 32 21 the the DET iajs-3467 32 22 application application NOUN iajs-3467 32 23 to to ADP iajs-3467 32 24 heart heart NOUN iajs-3467 32 25 disease disease NOUN iajs-3467 32 26 are be AUX iajs-3467 32 27 described describe VERB iajs-3467 32 28 in in ADP iajs-3467 32 29 sections section NOUN iajs-3467 32 30 6 6 NUM iajs-3467 32 31 and and CCONJ iajs-3467 32 32 7 7 NUM iajs-3467 32 33 , , PUNCT iajs-3467 32 34 respectively respectively ADV iajs-3467 32 35 . . PUNCT iajs-3467 33 1 finally finally ADV iajs-3467 33 2 , , PUNCT iajs-3467 33 3 the the DET iajs-3467 33 4 results result NOUN iajs-3467 33 5 and and CCONJ iajs-3467 33 6 discussion discussion NOUN iajs-3467 33 7 are be AUX iajs-3467 33 8 discussed discuss VERB iajs-3467 33 9 in in ADP iajs-3467 33 10 section section NOUN iajs-3467 33 11 8 8 NUM iajs-3467 33 12 . . NOUN iajs-3467 34 1 2 2 NUM iajs-3467 34 2 . . X iajs-3467 34 3 hard hard ADJ iajs-3467 34 4 margin margin NOUN iajs-3467 34 5 assume assume VERB iajs-3467 34 6 that that SCONJ iajs-3467 34 7 two two NUM iajs-3467 34 8 categories category NOUN iajs-3467 34 9 are be AUX iajs-3467 34 10 linearly linearly ADV iajs-3467 34 11 separable separable ADJ iajs-3467 34 12 , , PUNCT iajs-3467 34 13 so so SCONJ iajs-3467 34 14 there there PRON iajs-3467 34 15 exists exist VERB iajs-3467 34 16 a a DET iajs-3467 34 17 hyperplane hyperplane NOUN iajs-3467 34 18 𝑓(π‘₯ 𝑓(π‘₯ NOUN iajs-3467 34 19 ) ) PUNCT iajs-3467 35 1 = = SYM iajs-3467 35 2 𝛽0 𝛽0 NOUN iajs-3467 35 3 + + X iajs-3467 35 4 𝛽𝑇 𝛽𝑇 NOUN iajs-3467 35 5 = = SYM iajs-3467 35 6 0 0 NUM iajs-3467 35 7 that that PRON iajs-3467 35 8 separates separate VERB iajs-3467 35 9 the the DET iajs-3467 35 10 categories category NOUN iajs-3467 35 11 . . PUNCT iajs-3467 36 1 our our PRON iajs-3467 36 2 task task NOUN iajs-3467 36 3 is be AUX iajs-3467 36 4 to to PART iajs-3467 36 5 find find VERB iajs-3467 36 6 out out ADP iajs-3467 36 7 estimates estimate NOUN iajs-3467 36 8 of of ADP iajs-3467 36 9 𝛽0 𝛽0 NOUN iajs-3467 36 10 and and CCONJ iajs-3467 36 11 𝛽. 𝛽. NOUN iajs-3467 36 12 suppose suppose VERB iajs-3467 36 13 ( ( PUNCT iajs-3467 36 14 π‘₯(1 π‘₯(1 NOUN iajs-3467 36 15 ) ) PUNCT iajs-3467 36 16 , , PUNCT iajs-3467 36 17 𝑦1 𝑦1 PROPN iajs-3467 36 18 ) ) PUNCT iajs-3467 36 19 , , PUNCT iajs-3467 36 20 ( ( PUNCT iajs-3467 36 21 π‘₯(2 π‘₯(2 PROPN iajs-3467 36 22 ) ) PUNCT iajs-3467 36 23 , , PUNCT iajs-3467 36 24 𝑦2 𝑦2 PROPN iajs-3467 36 25 ) ) PUNCT iajs-3467 36 26 , , PUNCT iajs-3467 36 27 . . PUNCT iajs-3467 36 28 . . PUNCT iajs-3467 36 29 . . PUNCT iajs-3467 37 1 ( ( PUNCT iajs-3467 37 2 π‘₯(𝑛 π‘₯(𝑛 PROPN iajs-3467 37 3 ) ) PUNCT iajs-3467 37 4 , , PUNCT iajs-3467 37 5 𝑦𝑛 𝑦𝑛 NOUN iajs-3467 37 6 ) ) PUNCT iajs-3467 37 7 are be AUX iajs-3467 37 8 𝑛 𝑛 DET iajs-3467 37 9 data data NOUN iajs-3467 37 10 points point NOUN iajs-3467 37 11 from from ADP iajs-3467 37 12 above above ADP iajs-3467 37 13 setting set VERB iajs-3467 37 14 . . PUNCT iajs-3467 38 1 consider consider VERB iajs-3467 38 2 any any DET iajs-3467 38 3 hyper hyper ADJ iajs-3467 38 4 plane plane NOUN iajs-3467 38 5 𝛽0 𝛽0 NOUN iajs-3467 38 6 + + CCONJ iajs-3467 38 7 𝛽𝑇 𝛽𝑇 NOUN iajs-3467 38 8 = = SYM iajs-3467 38 9 0 0 NUM iajs-3467 38 10 in in ADP iajs-3467 38 11 the the DET iajs-3467 38 12 x x NOUN iajs-3467 38 13 - - NOUN iajs-3467 38 14 space space NOUN iajs-3467 38 15 . . PUNCT iajs-3467 39 1 the the DET iajs-3467 39 2 perpendicular perpendicular ADJ iajs-3467 39 3 distance distance NOUN iajs-3467 39 4 of of ADP iajs-3467 39 5 the the DET iajs-3467 39 6 π‘–π‘‘β„Ž π‘–π‘‘β„Ž NOUN iajs-3467 39 7 covariate covariate ADJ iajs-3467 39 8 point point NOUN iajs-3467 39 9 π‘₯(𝑖 π‘₯(𝑖 PROPN iajs-3467 39 10 ) ) PUNCT iajs-3467 39 11 from from ADP iajs-3467 39 12 this this DET iajs-3467 39 13 line line NOUN iajs-3467 39 14 is be AUX iajs-3467 39 15 𝑑𝑖 𝑑𝑖 NOUN iajs-3467 39 16 = = PROPN iajs-3467 39 17 𝑦𝑖 𝑦𝑖 NUM iajs-3467 39 18 𝛽0+𝛽𝑇𝑋(𝑖 𝛽0+𝛽𝑇𝑋(𝑖 PROPN iajs-3467 39 19 ) ) PUNCT iajs-3467 39 20 ǁ𝛽ǁ ǁ𝛽ǁ ADV iajs-3467 39 21 , , PUNCT iajs-3467 39 22 and and CCONJ iajs-3467 39 23 for for ADP iajs-3467 39 24 all all DET iajs-3467 39 25 training training NOUN iajs-3467 39 26 data datum NOUN iajs-3467 39 27 points point NOUN iajs-3467 39 28 𝑦𝑖(𝛽0 𝑦𝑖(𝛽0 VERB iajs-3467 39 29 + + CCONJ iajs-3467 39 30 𝛽𝑖π‘₯ 𝛽𝑖π‘₯ NOUN iajs-3467 39 31 ) ) PUNCT iajs-3467 39 32 > > X iajs-3467 39 33 0 0 PUNCT iajs-3467 40 1 [ [ X iajs-3467 40 2 6 6 NUM iajs-3467 40 3 ] ] PUNCT iajs-3467 40 4 . . PUNCT iajs-3467 41 1 the the DET iajs-3467 41 2 minimum minimum NOUN iajs-3467 41 3 of of ADP iajs-3467 41 4 these these DET iajs-3467 41 5 distance distance NOUN iajs-3467 41 6 is be AUX iajs-3467 41 7 called call VERB iajs-3467 41 8 the the DET iajs-3467 41 9 margin margin NOUN iajs-3467 41 10 , , PUNCT iajs-3467 41 11 i.e. i.e. X iajs-3467 41 12 , , PUNCT iajs-3467 41 13 there there PRON iajs-3467 41 14 is be VERB iajs-3467 41 15 no no DET iajs-3467 41 16 data data NOUN iajs-3467 41 17 point point NOUN iajs-3467 41 18 within within ADP iajs-3467 41 19 this this DET iajs-3467 41 20 distance distance NOUN iajs-3467 41 21 on on ADP iajs-3467 41 22 either either DET iajs-3467 41 23 side side NOUN iajs-3467 41 24 of of ADP iajs-3467 41 25 this this DET iajs-3467 41 26 line line NOUN iajs-3467 41 27 with with ADP iajs-3467 41 28 respect respect NOUN iajs-3467 41 29 to to ADP iajs-3467 41 30 this this DET iajs-3467 41 31 training training NOUN iajs-3467 41 32 dataset dataset NOUN iajs-3467 41 33 [ [ X iajs-3467 41 34 7 7 NUM iajs-3467 41 35 ] ] PUNCT iajs-3467 41 36 . . PUNCT iajs-3467 42 1 the the DET iajs-3467 42 2 aim aim NOUN iajs-3467 42 3 is be AUX iajs-3467 42 4 to to PART iajs-3467 42 5 find find VERB iajs-3467 42 6 a a DET iajs-3467 42 7 line line NOUN iajs-3467 42 8 that that PRON iajs-3467 42 9 has have VERB iajs-3467 42 10 maximum maximum ADJ iajs-3467 42 11 margin margin NOUN iajs-3467 42 12 among among ADP iajs-3467 42 13 all all DET iajs-3467 42 14 candidate candidate NOUN iajs-3467 42 15 lines line NOUN iajs-3467 42 16 . . PUNCT iajs-3467 43 1 that that DET iajs-3467 43 2 line line NOUN iajs-3467 43 3 is be AUX iajs-3467 43 4 going go VERB iajs-3467 43 5 to to PART iajs-3467 43 6 be be AUX iajs-3467 43 7 our our PRON iajs-3467 43 8 estimate estimate NOUN iajs-3467 43 9 of of ADP iajs-3467 43 10 classifier classifier NOUN iajs-3467 43 11 𝑓(𝑋 𝑓(𝑋 PROPN iajs-3467 43 12 ) ) PUNCT iajs-3467 43 13 . . PUNCT iajs-3467 44 1 the the DET iajs-3467 44 2 intuition intuition NOUN iajs-3467 44 3 behind behind ADP iajs-3467 44 4 keeping keep VERB iajs-3467 44 5 the the DET iajs-3467 44 6 margin margin NOUN iajs-3467 44 7 maximum maximum NOUN iajs-3467 44 8 is be AUX iajs-3467 44 9 a a DET iajs-3467 44 10 supervised supervised ADJ iajs-3467 44 11 learning learning NOUN iajs-3467 44 12 procedure procedure NOUN iajs-3467 44 13 . . PUNCT iajs-3467 45 1 we we PRON iajs-3467 45 2 want want VERB iajs-3467 45 3 to to PART iajs-3467 45 4 ensure ensure VERB iajs-3467 45 5 that that SCONJ iajs-3467 45 6 , , PUNCT iajs-3467 45 7 for for ADP iajs-3467 45 8 new new ADJ iajs-3467 45 9 observations observation NOUN iajs-3467 45 10 ( ( PUNCT iajs-3467 45 11 test test NOUN iajs-3467 45 12 data datum NOUN iajs-3467 45 13 ) ) PUNCT iajs-3467 45 14 , , PUNCT iajs-3467 45 15 we we PRON iajs-3467 45 16 accurately accurately ADV iajs-3467 45 17 determine determine VERB iajs-3467 45 18 category category NOUN iajs-3467 45 19 of of ADP iajs-3467 45 20 𝑦. 𝑦. PROPN iajs-3467 45 21 we we PRON iajs-3467 45 22 want want VERB iajs-3467 45 23 to to PART iajs-3467 45 24 maximize maximize VERB iajs-3467 45 25 the the DET iajs-3467 45 26 margin margin NOUN iajs-3467 45 27 because because SCONJ iajs-3467 45 28 it it PRON iajs-3467 45 29 gives give VERB iajs-3467 45 30 us we PRON iajs-3467 45 31 a a DET iajs-3467 45 32 room room NOUN iajs-3467 45 33 for for ADP iajs-3467 45 34 allowing allow VERB iajs-3467 45 35 for for ADP iajs-3467 45 36 variation variation NOUN iajs-3467 45 37 between between ADP iajs-3467 45 38 training training NOUN iajs-3467 45 39 and and CCONJ iajs-3467 45 40 test test NOUN iajs-3467 45 41 datasets dataset NOUN iajs-3467 45 42 . . PUNCT iajs-3467 46 1 assume assume VERB iajs-3467 46 2 the the DET iajs-3467 46 3 optimization optimization NOUN iajs-3467 46 4 problem problem NOUN iajs-3467 46 5 is be AUX iajs-3467 46 6 π‘šπ‘–π‘›π›½0,𝛽 π‘šπ‘–π‘›π›½0,𝛽 PROPN iajs-3467 46 7 ǁ𝛽ǁ2 ǁ𝛽ǁ2 ADJ iajs-3467 46 8 2 2 NUM iajs-3467 46 9 subject subject NOUN iajs-3467 46 10 to to ADP iajs-3467 46 11 𝑦𝑖(𝛽0 𝑦𝑖(𝛽0 NOUN iajs-3467 46 12 + + ADV iajs-3467 46 13 𝛽𝑇π‘₯(𝑖 𝛽𝑇π‘₯(𝑖 NUM iajs-3467 46 14 ) ) PUNCT iajs-3467 46 15 ) ) PUNCT iajs-3467 47 1 β‰₯ β‰₯ NOUN iajs-3467 47 2 1 1 NUM iajs-3467 47 3 π‘“π‘œπ‘Ÿ π‘“π‘œπ‘Ÿ NOUN iajs-3467 47 4 𝑖 𝑖 NOUN iajs-3467 47 5 = = SYM iajs-3467 47 6 1,2 1,2 NUM iajs-3467 47 7 , , PUNCT iajs-3467 47 8 . . PUNCT iajs-3467 47 9 . . PUNCT iajs-3467 48 1 . . PUNCT iajs-3467 49 1 , , PUNCT iajs-3467 49 2 𝑛 𝑛 X iajs-3467 49 3 ( ( PUNCT iajs-3467 49 4 1 1 NUM iajs-3467 49 5 ) ) PUNCT iajs-3467 49 6 if if SCONJ iajs-3467 49 7 ( ( PUNCT iajs-3467 49 8 οΏ½ οΏ½ NOUN iajs-3467 49 9 Μ‚ Μ‚ VERB iajs-3467 49 10 οΏ½ οΏ½ NOUN iajs-3467 49 11 0 0 NUM iajs-3467 49 12 , , PUNCT iajs-3467 49 13 οΏ½ οΏ½ NOUN iajs-3467 49 14 Μ‚ Μ‚ NOUN iajs-3467 49 15 οΏ½ οΏ½ NOUN iajs-3467 49 16 ) ) PUNCT iajs-3467 49 17 represent represent VERB iajs-3467 49 18 the the DET iajs-3467 49 19 solution solution NOUN iajs-3467 49 20 to to ADP iajs-3467 49 21 this this DET iajs-3467 49 22 optimization optimization NOUN iajs-3467 49 23 problem problem NOUN iajs-3467 49 24 , , PUNCT iajs-3467 49 25 some some DET iajs-3467 49 26 statements statement NOUN iajs-3467 49 27 according accord VERB iajs-3467 49 28 to to ADP iajs-3467 49 29 eq(1 eq(1 NOUN iajs-3467 49 30 ) ) PUNCT iajs-3467 49 31 can can AUX iajs-3467 49 32 be be AUX iajs-3467 49 33 hold hold VERB iajs-3467 49 34 . . PUNCT iajs-3467 50 1 first first ADV iajs-3467 50 2 , , PUNCT iajs-3467 50 3 Ζ’Μ‚(π‘₯ Ζ’Μ‚(π‘₯ NOUN iajs-3467 50 4 ) ) PUNCT iajs-3467 50 5 = = SYM iajs-3467 50 6 οΏ½ οΏ½ PROPN iajs-3467 50 7 Μ‚ Μ‚ NOUN iajs-3467 50 8 οΏ½ οΏ½ NOUN iajs-3467 50 9 0 0 NUM iajs-3467 50 10 + + NUM iajs-3467 50 11 οΏ½ οΏ½ PROPN iajs-3467 50 12 Μ‚ Μ‚ NOUN iajs-3467 50 13 οΏ½ οΏ½ NOUN iajs-3467 50 14 𝑇𝑋 𝑇𝑋 NOUN iajs-3467 50 15 = = SYM iajs-3467 50 16 0 0 NUM iajs-3467 50 17 represents represent VERB iajs-3467 50 18 the the DET iajs-3467 50 19 estimated estimate VERB iajs-3467 50 20 linear linear PROPN iajs-3467 50 21 classifier classifier NOUN iajs-3467 50 22 . . PUNCT iajs-3467 51 1 for for ADP iajs-3467 51 2 any any DET iajs-3467 51 3 new new ADJ iajs-3467 51 4 data datum NOUN iajs-3467 51 5 point point NOUN iajs-3467 51 6 with with ADP iajs-3467 51 7 covariate covariate ADJ iajs-3467 51 8 information information NOUN iajs-3467 51 9 π‘₯0 π‘₯0 NOUN iajs-3467 51 10 , , PUNCT iajs-3467 51 11 evaluate evaluate VERB iajs-3467 51 12 Ζ’Μ‚(π‘₯0 Ζ’Μ‚(π‘₯0 NOUN iajs-3467 51 13 ) ) PUNCT iajs-3467 51 14 = = SYM iajs-3467 51 15 οΏ½ οΏ½ PROPN iajs-3467 51 16 Μ‚ Μ‚ NOUN iajs-3467 51 17 οΏ½ οΏ½ NOUN iajs-3467 51 18 0 0 NUM iajs-3467 51 19 + + NUM iajs-3467 51 20 οΏ½ οΏ½ PROPN iajs-3467 51 21 Μ‚ Μ‚ NUM iajs-3467 51 22 οΏ½ οΏ½ NOUN iajs-3467 51 23 𝑇𝑋0 𝑇𝑋0 PROPN iajs-3467 51 24 . . PUNCT iajs-3467 52 1 if if SCONJ iajs-3467 52 2 Ζ’Μ‚(π‘₯0 Ζ’Μ‚(π‘₯0 PROPN iajs-3467 52 3 ) ) PUNCT iajs-3467 52 4 > > X iajs-3467 52 5 0 0 NUM iajs-3467 52 6 , , PUNCT iajs-3467 52 7 decide decide VERB iajs-3467 52 8 category1 category1 NOUN iajs-3467 52 9 for for ADP iajs-3467 52 10 that that DET iajs-3467 52 11 ihjpas ihjpa NOUN iajs-3467 52 12 . . PUNCT iajs-3467 53 1 37 37 NUM iajs-3467 53 2 ( ( PUNCT iajs-3467 53 3 1 1 NUM iajs-3467 53 4 ) ) PUNCT iajs-3467 53 5 2024 2024 NUM iajs-3467 53 6 414 414 NUM iajs-3467 53 7 observation observation NOUN iajs-3467 53 8 , , PUNCT iajs-3467 53 9 otherwise otherwise ADV iajs-3467 53 10 if if SCONJ iajs-3467 53 11 Ζ’Μ‚(π‘₯0 Ζ’Μ‚(π‘₯0 X iajs-3467 53 12 ) ) PUNCT iajs-3467 53 13 < < X iajs-3467 53 14 0 0 NUM iajs-3467 53 15 , , PUNCT iajs-3467 53 16 decide decide VERB iajs-3467 53 17 category category NOUN iajs-3467 53 18 2 2 NUM iajs-3467 53 19 . . PUNCT iajs-3467 53 20 second second ADJ iajs-3467 53 21 , , PUNCT iajs-3467 53 22 the the DET iajs-3467 53 23 margin margin NOUN iajs-3467 53 24 of of ADP iajs-3467 53 25 the the DET iajs-3467 53 26 classifier classifier NOUN iajs-3467 53 27 is be AUX iajs-3467 53 28 : : PUNCT iajs-3467 53 29 1 1 NUM iajs-3467 53 30 ǁ ǁ X iajs-3467 53 31 οΏ½ οΏ½ PROPN iajs-3467 53 32 Μ‚ Μ‚ SYM iajs-3467 53 33 οΏ½ οΏ½ PROPN iajs-3467 53 34 ǁ ǁ NOUN iajs-3467 53 35 . . PUNCT iajs-3467 54 1 geometrically geometrically ADV iajs-3467 54 2 , , PUNCT iajs-3467 54 3 there there PRON iajs-3467 54 4 will will AUX iajs-3467 54 5 not not PART iajs-3467 54 6 be be AUX iajs-3467 54 7 any any DET iajs-3467 54 8 training training NOUN iajs-3467 54 9 data datum NOUN iajs-3467 54 10 points point NOUN iajs-3467 54 11 between between ADP iajs-3467 54 12 the the DET iajs-3467 54 13 lines line NOUN iajs-3467 54 14 οΏ½ οΏ½ NOUN iajs-3467 54 15 Μ‚ Μ‚ VERB iajs-3467 54 16 οΏ½ οΏ½ NOUN iajs-3467 54 17 0 0 NUM iajs-3467 54 18 + + NUM iajs-3467 54 19 οΏ½ οΏ½ PROPN iajs-3467 54 20 Μ‚ Μ‚ NOUN iajs-3467 54 21 οΏ½ οΏ½ NOUN iajs-3467 54 22 𝑇𝑋 𝑇𝑋 NOUN iajs-3467 54 23 = = SYM iajs-3467 54 24 +1 +1 PROPN iajs-3467 54 25 and and CCONJ iajs-3467 54 26 οΏ½ οΏ½ PROPN iajs-3467 54 27 Μ‚ Μ‚ NOUN iajs-3467 54 28 οΏ½ οΏ½ NOUN iajs-3467 54 29 0 0 NUM iajs-3467 54 30 + + NUM iajs-3467 54 31 οΏ½ οΏ½ PROPN iajs-3467 54 32 Μ‚ Μ‚ NOUN iajs-3467 54 33 οΏ½ οΏ½ NOUN iajs-3467 54 34 𝑇𝑋 𝑇𝑋 NOUN iajs-3467 54 35 = = SYM iajs-3467 54 36 βˆ’1 βˆ’1 NOUN iajs-3467 54 37 . . PUNCT iajs-3467 55 1 as as SCONJ iajs-3467 55 2 shown show VERB iajs-3467 55 3 in in ADP iajs-3467 55 4 fig.1 fig.1 PROPN iajs-3467 55 5 , , PUNCT iajs-3467 55 6 many many ADJ iajs-3467 55 7 lines line NOUN iajs-3467 55 8 can can AUX iajs-3467 55 9 separate separate VERB iajs-3467 55 10 two two NUM iajs-3467 55 11 groups group NOUN iajs-3467 55 12 ( ( PUNCT iajs-3467 55 13 𝐻1 𝐻1 NOUN iajs-3467 55 14 , , PUNCT iajs-3467 55 15 𝐻2 𝐻2 ADJ iajs-3467 55 16 , , PUNCT iajs-3467 55 17 𝐻3 𝐻3 NOUN iajs-3467 55 18 , , PUNCT iajs-3467 55 19 and and CCONJ iajs-3467 55 20 𝐻4 𝐻4 PROPN iajs-3467 55 21 ) ) PUNCT iajs-3467 55 22 , , PUNCT iajs-3467 55 23 but but CCONJ iajs-3467 55 24 there there PRON iajs-3467 55 25 is be VERB iajs-3467 55 26 only only ADV iajs-3467 55 27 one one NUM iajs-3467 55 28 optimal optimal ADJ iajs-3467 55 29 separating separate VERB iajs-3467 55 30 hyperplane hyperplane NOUN iajs-3467 55 31 ( ( PUNCT iajs-3467 55 32 𝐻 𝐻 PROPN iajs-3467 55 33 ) ) PUNCT iajs-3467 55 34 with with ADP iajs-3467 55 35 two two NUM iajs-3467 55 36 boundary boundary ADJ iajs-3467 55 37 hyperplanes hyperplane NOUN iajs-3467 55 38 ( ( PUNCT iajs-3467 55 39 π»π‘š1 π»π‘š1 NOUN iajs-3467 55 40 π‘Žπ‘›π‘‘ π‘Žπ‘›π‘‘ PROPN iajs-3467 55 41 π»π‘š2 π»π‘š2 NOUN iajs-3467 55 42 ) ) PUNCT iajs-3467 55 43 see see VERB iajs-3467 55 44 fig.2 fig.2 PROPN iajs-3467 55 45 . . PUNCT iajs-3467 55 46 third third ADJ iajs-3467 55 47 , , PUNCT iajs-3467 55 48 the the DET iajs-3467 55 49 training training NOUN iajs-3467 55 50 data datum NOUN iajs-3467 55 51 points point NOUN iajs-3467 55 52 lie lie VERB iajs-3467 55 53 exactly exactly ADV iajs-3467 55 54 on on ADP iajs-3467 55 55 one one NUM iajs-3467 55 56 of of ADP iajs-3467 55 57 the the DET iajs-3467 55 58 above above ADJ iajs-3467 55 59 two two NUM iajs-3467 55 60 lines line NOUN iajs-3467 55 61 ( ( PUNCT iajs-3467 55 62 margins margin NOUN iajs-3467 55 63 ) ) PUNCT iajs-3467 55 64 are be AUX iajs-3467 55 65 referred refer VERB iajs-3467 55 66 to to ADP iajs-3467 55 67 as as ADP iajs-3467 55 68 support support NOUN iajs-3467 55 69 vectors vector NOUN iajs-3467 55 70 , , PUNCT iajs-3467 55 71 so so SCONJ iajs-3467 55 72 the the DET iajs-3467 55 73 estimates estimate NOUN iajs-3467 55 74 οΏ½ οΏ½ NOUN iajs-3467 55 75 Μ‚ Μ‚ VERB iajs-3467 55 76 οΏ½ οΏ½ NOUN iajs-3467 55 77 0 0 NUM iajs-3467 55 78 , , PUNCT iajs-3467 55 79 𝛽 𝛽 NOUN iajs-3467 55 80 depend depend VERB iajs-3467 55 81 only only ADV iajs-3467 55 82 on on ADP iajs-3467 55 83 the the DET iajs-3467 55 84 data data NOUN iajs-3467 55 85 points point NOUN iajs-3467 55 86 that that PRON iajs-3467 55 87 are be AUX iajs-3467 55 88 support support NOUN iajs-3467 55 89 vectors vector NOUN iajs-3467 55 90 . . PUNCT iajs-3467 56 1 that that PRON iajs-3467 56 2 means mean VERB iajs-3467 56 3 all all DET iajs-3467 56 4 data datum NOUN iajs-3467 56 5 points point NOUN iajs-3467 56 6 inside inside ADP iajs-3467 56 7 the the DET iajs-3467 56 8 correct correct ADJ iajs-3467 56 9 margins margin NOUN iajs-3467 56 10 have have VERB iajs-3467 56 11 no no DET iajs-3467 56 12 role role NOUN iajs-3467 56 13 in in ADP iajs-3467 56 14 determining determine VERB iajs-3467 56 15 the the DET iajs-3467 56 16 form form NOUN iajs-3467 56 17 of of ADP iajs-3467 56 18 the the DET iajs-3467 56 19 classifier classifier NOUN iajs-3467 56 20 . . PUNCT iajs-3467 57 1 this this DET iajs-3467 57 2 property property NOUN iajs-3467 57 3 makes make VERB iajs-3467 57 4 svm svm ADJ iajs-3467 57 5 very very ADV iajs-3467 57 6 useful useful ADJ iajs-3467 57 7 for for ADP iajs-3467 57 8 massive massive ADJ iajs-3467 57 9 data data NOUN iajs-3467 57 10 classification classification NOUN iajs-3467 57 11 problems problem NOUN iajs-3467 57 12 [ [ X iajs-3467 57 13 7 7 NUM iajs-3467 57 14 ] ] PUNCT iajs-3467 57 15 . . PUNCT iajs-3467 58 1 figure figure NOUN iajs-3467 58 2 1 1 NUM iajs-3467 58 3 . . PUNCT iajs-3467 59 1 some some DET iajs-3467 59 2 lines line NOUN iajs-3467 59 3 that that PRON iajs-3467 59 4 separate separate VERB iajs-3467 59 5 two two NUM iajs-3467 59 6 groups group NOUN iajs-3467 59 7 : : PUNCT iajs-3467 59 8 𝐻1 𝐻1 NOUN iajs-3467 59 9 , , PUNCT iajs-3467 59 10 𝐻2 𝐻2 ADJ iajs-3467 59 11 , , PUNCT iajs-3467 59 12 𝐻3 𝐻3 PROPN iajs-3467 59 13 , , PUNCT iajs-3467 59 14 π‘Žπ‘›π‘‘ π‘Žπ‘›π‘‘ ADJ iajs-3467 59 15 𝐻4 𝐻4 PROPN iajs-3467 59 16 . . PUNCT iajs-3467 59 17 ihjpas ihjpas PROPN iajs-3467 59 18 . . PUNCT iajs-3467 60 1 37 37 NUM iajs-3467 60 2 ( ( PUNCT iajs-3467 60 3 1 1 NUM iajs-3467 60 4 ) ) PUNCT iajs-3467 60 5 2024 2024 NUM iajs-3467 60 6 415 415 NUM iajs-3467 60 7 figure figure NOUN iajs-3467 60 8 2 2 NUM iajs-3467 60 9 . . X iajs-3467 60 10 separating separate VERB iajs-3467 60 11 hyperplane hyperplane NOUN iajs-3467 60 12 , , PUNCT iajs-3467 60 13 𝐻 𝐻 PROPN iajs-3467 60 14 , , PUNCT iajs-3467 60 15 with with ADP iajs-3467 60 16 two two NUM iajs-3467 60 17 boundary boundary ADJ iajs-3467 60 18 hyperplanes hyperplane NOUN iajs-3467 60 19 : : PUNCT iajs-3467 60 20 π»π‘š1 π»π‘š1 VERB iajs-3467 60 21 π‘Žπ‘›π‘‘ π‘Žπ‘›π‘‘ PROPN iajs-3467 60 22 π»π‘š2 π»π‘š2 NOUN iajs-3467 60 23 . . PUNCT iajs-3467 61 1 3 3 X iajs-3467 61 2 . . X iajs-3467 61 3 soft soft ADJ iajs-3467 61 4 margin margin NOUN iajs-3467 61 5 suppose suppose VERB iajs-3467 61 6 the the DET iajs-3467 61 7 assumption assumption NOUN iajs-3467 61 8 of of ADP iajs-3467 61 9 linear linear PROPN iajs-3467 61 10 separability separability NOUN iajs-3467 61 11 does do AUX iajs-3467 61 12 not not PART iajs-3467 61 13 hold hold VERB iajs-3467 61 14 between between ADP iajs-3467 61 15 the the DET iajs-3467 61 16 two two NUM iajs-3467 61 17 categories category NOUN iajs-3467 61 18 . . PUNCT iajs-3467 62 1 that that PRON iajs-3467 62 2 implies imply VERB iajs-3467 62 3 , , PUNCT iajs-3467 62 4 no no ADV iajs-3467 62 5 matter matter ADV iajs-3467 62 6 what what PRON iajs-3467 62 7 hyper hyper ADJ iajs-3467 62 8 plane plane NOUN iajs-3467 62 9 𝛽0 𝛽0 NOUN iajs-3467 62 10 + + CCONJ iajs-3467 62 11 𝛽𝑇𝑋 𝛽𝑇𝑋 ADJ iajs-3467 62 12 = = SYM iajs-3467 62 13 0 0 NUM iajs-3467 62 14 , , PUNCT iajs-3467 62 15 there there PRON iajs-3467 62 16 will will AUX iajs-3467 62 17 always always ADV iajs-3467 62 18 be be AUX iajs-3467 62 19 points point NOUN iajs-3467 62 20 ( ( PUNCT iajs-3467 62 21 π‘₯ π‘₯ NOUN iajs-3467 62 22 , , PUNCT iajs-3467 62 23 𝑦 𝑦 NOUN iajs-3467 62 24 ) ) PUNCT iajs-3467 62 25 such such ADJ iajs-3467 62 26 that that SCONJ iajs-3467 62 27 : : PUNCT iajs-3467 62 28 𝛽0 𝛽0 NOUN iajs-3467 62 29 + + CCONJ iajs-3467 62 30 𝛽𝑇π‘₯ 𝛽𝑇π‘₯ NOUN iajs-3467 62 31 > > SYM iajs-3467 62 32 0 0 NUM iajs-3467 62 33 𝑏𝑒𝑑 𝑏𝑒𝑑 NOUN iajs-3467 62 34 𝑦 𝑦 NOUN iajs-3467 62 35 = = X iajs-3467 62 36 βˆ’1 βˆ’1 NOUN iajs-3467 62 37 π‘Žπ‘›π‘‘ π‘Žπ‘›π‘‘ PROPN iajs-3467 62 38 𝛽0 𝛽0 NOUN iajs-3467 62 39 + + CCONJ iajs-3467 62 40 𝛽𝑇π‘₯ 𝛽𝑇π‘₯ VERB iajs-3467 62 41 < < X iajs-3467 62 42 0 0 NUM iajs-3467 62 43 𝑏𝑒𝑑 𝑏𝑒𝑑 NOUN iajs-3467 62 44 𝑦 𝑦 PROPN iajs-3467 62 45 = = X iajs-3467 62 46 +1 +1 PROPN iajs-3467 62 47 . . PUNCT iajs-3467 62 48 consider consider VERB iajs-3467 62 49 a a DET iajs-3467 62 50 nonnegative nonnegative ADJ iajs-3467 62 51 variable variable NOUN iajs-3467 62 52 𝐾𝑖 𝐾𝑖 PROPN iajs-3467 62 53 for for ADP iajs-3467 62 54 each each DET iajs-3467 62 55 observation observation NOUN iajs-3467 62 56 i i NOUN iajs-3467 62 57 = = NOUN iajs-3467 62 58 1,2 1,2 NUM iajs-3467 62 59 , , PUNCT iajs-3467 62 60 … … PUNCT iajs-3467 62 61 , , PUNCT iajs-3467 62 62 n. n. NOUN iajs-3467 63 1 𝐾𝑖 𝐾𝑖 PROPN iajs-3467 63 2 denotes denote VERB iajs-3467 63 3 the the DET iajs-3467 63 4 amount amount NOUN iajs-3467 63 5 of of ADP iajs-3467 63 6 push push NOUN iajs-3467 63 7 an an DET iajs-3467 63 8 observation observation NOUN iajs-3467 63 9 needs need VERB iajs-3467 63 10 to to PART iajs-3467 63 11 go go VERB iajs-3467 63 12 the the DET iajs-3467 63 13 correct correct ADJ iajs-3467 63 14 side side NOUN iajs-3467 63 15 of of ADP iajs-3467 63 16 the the DET iajs-3467 63 17 margin margin NOUN iajs-3467 63 18 . . PUNCT iajs-3467 64 1 we we PRON iajs-3467 64 2 have have AUX iajs-3467 64 3 shown show VERB iajs-3467 64 4 in in ADP iajs-3467 64 5 class class NOUN iajs-3467 64 6 that that SCONJ iajs-3467 64 7 [ [ X iajs-3467 64 8 8 8 NUM iajs-3467 64 9 ] ] PUNCT iajs-3467 64 10 : : PUNCT iajs-3467 64 11 a. a. NOUN iajs-3467 64 12 for for ADP iajs-3467 64 13 points point NOUN iajs-3467 64 14 already already ADV iajs-3467 64 15 obeying obey VERB iajs-3467 64 16 correct correct ADJ iajs-3467 64 17 margin margin NOUN iajs-3467 64 18 , , PUNCT iajs-3467 64 19 𝐾𝑖 𝐾𝑖 PROPN iajs-3467 64 20 = = SYM iajs-3467 64 21 0 0 X iajs-3467 64 22 . . X iajs-3467 64 23 b. b. PROPN iajs-3467 64 24 for for ADP iajs-3467 64 25 points point NOUN iajs-3467 64 26 violating violate VERB iajs-3467 64 27 the the DET iajs-3467 64 28 margin margin NOUN iajs-3467 64 29 but but CCONJ iajs-3467 64 30 staying stay VERB iajs-3467 64 31 on on ADP iajs-3467 64 32 the the DET iajs-3467 64 33 correct correct ADJ iajs-3467 64 34 side side NOUN iajs-3467 64 35 of of ADP iajs-3467 64 36 the the DET iajs-3467 64 37 line line NOUN iajs-3467 64 38 , , PUNCT iajs-3467 64 39 0 0 PUNCT iajs-3467 64 40 < < X iajs-3467 65 1 𝐾𝑖 𝐾𝑖 PUNCT iajs-3467 65 2 < < X iajs-3467 65 3 1 1 NUM iajs-3467 65 4 ǁ𝛽ǁ ǁ𝛽ǁ NOUN iajs-3467 65 5 . . PUNCT iajs-3467 66 1 c. c. NOUN iajs-3467 66 2 for for ADP iajs-3467 66 3 points point NOUN iajs-3467 66 4 violating violate VERB iajs-3467 66 5 the the DET iajs-3467 66 6 margin margin NOUN iajs-3467 66 7 and and CCONJ iajs-3467 66 8 moving move VERB iajs-3467 66 9 to to ADP iajs-3467 66 10 the the DET iajs-3467 66 11 other other ADJ iajs-3467 66 12 side side NOUN iajs-3467 66 13 of of ADP iajs-3467 66 14 the the DET iajs-3467 66 15 line line NOUN iajs-3467 67 1 𝐾𝑖 𝐾𝑖 INTJ iajs-3467 67 2 > > SYM iajs-3467 67 3 1 1 NUM iajs-3467 67 4 ǁ𝛽ǁ ǁ𝛽ǁ NOUN iajs-3467 67 5 . . PUNCT iajs-3467 68 1 by by ADP iajs-3467 68 2 solving solve VERB iajs-3467 68 3 the the DET iajs-3467 68 4 optimization optimization NOUN iajs-3467 68 5 problem problem NOUN iajs-3467 68 6 , , PUNCT iajs-3467 68 7 the the DET iajs-3467 68 8 estimates estimate NOUN iajs-3467 68 9 of of ADP iajs-3467 68 10 ( ( PUNCT iajs-3467 68 11 𝛽0 𝛽0 PROPN iajs-3467 68 12 , , PUNCT iajs-3467 68 13 𝛽 𝛽 NOUN iajs-3467 68 14 ) ) PUNCT iajs-3467 68 15 can can AUX iajs-3467 68 16 be be AUX iajs-3467 68 17 found find VERB iajs-3467 68 18 : : PUNCT iajs-3467 68 19 π‘šπ‘–π‘› π‘šπ‘–π‘› PROPN iajs-3467 68 20 𝛽0,𝛽 𝛽0,𝛽 PROPN iajs-3467 68 21 , , PUNCT iajs-3467 68 22 𝑠𝑖 𝑠𝑖 NOUN iajs-3467 68 23 ǁ𝛽ǁ2 ǁ𝛽ǁ2 NOUN iajs-3467 68 24 2 2 NUM iajs-3467 69 1 + + CCONJ iajs-3467 69 2 𝐢 𝐢 PROPN iajs-3467 69 3 βˆ‘ βˆ‘ PUNCT iajs-3467 69 4 𝑠𝑖 𝑠𝑖 NOUN iajs-3467 69 5 𝑛 𝑛 PRON iajs-3467 69 6 𝑖=1 𝑖=1 PROPN iajs-3467 69 7 𝑠𝑒𝑏𝑗𝑒𝑐𝑑 𝑠𝑒𝑏𝑗𝑒𝑐𝑑 NOUN iajs-3467 69 8 π‘‘π‘œ π‘‘π‘œ PROPN iajs-3467 69 9 𝑦𝑖(𝛽0 𝑦𝑖(𝛽0 VERB iajs-3467 69 10 + + CCONJ iajs-3467 69 11 𝛽𝑇𝑋(𝑖 𝛽𝑇𝑋(𝑖 NOUN iajs-3467 69 12 ) ) PUNCT iajs-3467 69 13 ) ) PUNCT iajs-3467 70 1 β‰₯ β‰₯ NOUN iajs-3467 70 2 1 1 NUM iajs-3467 70 3 βˆ’ βˆ’ NOUN iajs-3467 70 4 𝑠𝑖 𝑠𝑖 NOUN iajs-3467 70 5 , , PUNCT iajs-3467 70 6 𝑠𝑖 𝑠𝑖 VERB iajs-3467 70 7 β‰₯ β‰₯ NOUN iajs-3467 70 8 0 0 NUM iajs-3467 71 1 π‘“π‘œπ‘Ÿ π‘“π‘œπ‘Ÿ NOUN iajs-3467 71 2 𝑖 𝑖 X iajs-3467 71 3 = = SYM iajs-3467 71 4 1,2 1,2 NUM iajs-3467 71 5 , , PUNCT iajs-3467 71 6 . . PUNCT iajs-3467 71 7 . . PUNCT iajs-3467 72 1 . . PUNCT iajs-3467 73 1 , , PUNCT iajs-3467 73 2 𝑛 𝑛 X iajs-3467 73 3 ( ( PUNCT iajs-3467 73 4 2 2 NUM iajs-3467 73 5 ) ) PUNCT iajs-3467 73 6 in in ADP iajs-3467 73 7 eq eq NOUN iajs-3467 73 8 ( ( PUNCT iajs-3467 73 9 2 2 NUM iajs-3467 73 10 ) ) PUNCT iajs-3467 73 11 , , PUNCT iajs-3467 73 12 c c PROPN iajs-3467 73 13 is be AUX iajs-3467 73 14 a a DET iajs-3467 73 15 pre pre ADJ iajs-3467 73 16 - - ADJ iajs-3467 73 17 fixed fixed ADJ iajs-3467 73 18 turning turning NOUN iajs-3467 73 19 parameter parameter NOUN iajs-3467 73 20 that that PRON iajs-3467 73 21 balances balance VERB iajs-3467 73 22 the the DET iajs-3467 73 23 relative relative ADJ iajs-3467 73 24 importance importance NOUN iajs-3467 73 25 of of ADP iajs-3467 73 26 maximizing maximize VERB iajs-3467 73 27 the the DET iajs-3467 73 28 margin margin NOUN iajs-3467 73 29 and and CCONJ iajs-3467 73 30 minimizing minimize VERB iajs-3467 73 31 the the DET iajs-3467 73 32 total total ADJ iajs-3467 73 33 amount amount NOUN iajs-3467 73 34 of of ADP iajs-3467 73 35 push push NOUN iajs-3467 73 36 required require VERB iajs-3467 73 37 for for ADP iajs-3467 73 38 points point NOUN iajs-3467 73 39 violating violate VERB iajs-3467 73 40 the the DET iajs-3467 73 41 margin margin NOUN iajs-3467 73 42 . . PUNCT iajs-3467 74 1 if if SCONJ iajs-3467 74 2 ( ( PUNCT iajs-3467 74 3 οΏ½ οΏ½ NOUN iajs-3467 74 4 Μ‚ Μ‚ VERB iajs-3467 74 5 οΏ½ οΏ½ NOUN iajs-3467 74 6 0 0 NUM iajs-3467 74 7 , , PUNCT iajs-3467 74 8 𝛽 𝛽 NOUN iajs-3467 74 9 ) ) PUNCT iajs-3467 74 10 represent represent VERB iajs-3467 74 11 the the DET iajs-3467 74 12 solution solution NOUN iajs-3467 74 13 to to ADP iajs-3467 74 14 this this DET iajs-3467 74 15 optimization optimization NOUN iajs-3467 74 16 problem problem NOUN iajs-3467 74 17 , , PUNCT iajs-3467 74 18 then then ADV iajs-3467 74 19 the the DET iajs-3467 74 20 following following ADJ iajs-3467 74 21 statements statement NOUN iajs-3467 74 22 hold hold VERB iajs-3467 74 23 : : PUNCT iajs-3467 74 24 1 1 X iajs-3467 74 25 . . PUNCT iajs-3467 74 26 𝑓(π‘₯ 𝑓(π‘₯ NOUN iajs-3467 74 27 ) ) PUNCT iajs-3467 74 28 = = SYM iajs-3467 74 29 οΏ½ οΏ½ PROPN iajs-3467 74 30 Μ‚ Μ‚ NOUN iajs-3467 74 31 οΏ½ οΏ½ NOUN iajs-3467 74 32 0 0 NUM iajs-3467 74 33 + + NUM iajs-3467 74 34 οΏ½ οΏ½ PROPN iajs-3467 74 35 Μ‚ Μ‚ NOUN iajs-3467 74 36 οΏ½ οΏ½ NOUN iajs-3467 74 37 𝑇π‘₯ 𝑇π‘₯ NOUN iajs-3467 74 38 = = SYM iajs-3467 74 39 0 0 NUM iajs-3467 74 40 represents represent VERB iajs-3467 74 41 the the DET iajs-3467 74 42 estimated estimate VERB iajs-3467 74 43 linear linear PROPN iajs-3467 74 44 classifier classifier NOUN iajs-3467 74 45 . . PUNCT iajs-3467 75 1 for for ADP iajs-3467 75 2 any any DET iajs-3467 75 3 new new ADJ iajs-3467 75 4 data datum NOUN iajs-3467 75 5 point point NOUN iajs-3467 75 6 with with ADP iajs-3467 75 7 covariate covariate ADJ iajs-3467 75 8 information information NOUN iajs-3467 75 9 π‘₯0 π‘₯0 NOUN iajs-3467 75 10 , , PUNCT iajs-3467 75 11 evaluate evaluate VERB iajs-3467 75 12 𝑓(π‘₯0 𝑓(π‘₯0 NOUN iajs-3467 75 13 ) ) PUNCT iajs-3467 75 14 = = PUNCT iajs-3467 75 15 οΏ½ οΏ½ PROPN iajs-3467 75 16 Μ‚ Μ‚ NOUN iajs-3467 75 17 οΏ½ οΏ½ NOUN iajs-3467 75 18 0 0 NUM iajs-3467 75 19 + + NUM iajs-3467 75 20 οΏ½ οΏ½ PROPN iajs-3467 75 21 Μ‚ Μ‚ NOUN iajs-3467 75 22 οΏ½ οΏ½ NOUN iajs-3467 75 23 𝑇π‘₯0 𝑇π‘₯0 NOUN iajs-3467 75 24 . . PUNCT iajs-3467 76 1 𝐼𝑓 𝐼𝑓 ADJ iajs-3467 76 2 𝑓(π‘₯0 𝑓(π‘₯0 NOUN iajs-3467 76 3 ) ) PUNCT iajs-3467 76 4 > > X iajs-3467 76 5 0 0 NUM iajs-3467 76 6 , , PUNCT iajs-3467 76 7 decide decide VERB iajs-3467 76 8 category category NOUN iajs-3467 76 9 1 1 NUM iajs-3467 76 10 for for ADP iajs-3467 76 11 that that DET iajs-3467 76 12 observation observation NOUN iajs-3467 76 13 , , PUNCT iajs-3467 76 14 otherwise otherwise ADV iajs-3467 76 15 if if SCONJ iajs-3467 76 16 𝑓(π‘₯0 𝑓(π‘₯0 NUM iajs-3467 76 17 ) ) PUNCT iajs-3467 76 18 < < X iajs-3467 76 19 0 0 NUM iajs-3467 76 20 , , PUNCT iajs-3467 76 21 decide decide VERB iajs-3467 76 22 category category NOUN iajs-3467 76 23 2 2 NUM iajs-3467 76 24 . . PUNCT iajs-3467 76 25 ihjpas ihjpas PROPN iajs-3467 76 26 . . PUNCT iajs-3467 77 1 37 37 NUM iajs-3467 77 2 ( ( PUNCT iajs-3467 77 3 1 1 NUM iajs-3467 77 4 ) ) PUNCT iajs-3467 77 5 2024 2024 NUM iajs-3467 77 6 416 416 NUM iajs-3467 77 7 2.the 2.the DET iajs-3467 77 8 margin margin NOUN iajs-3467 77 9 of of ADP iajs-3467 77 10 the the DET iajs-3467 77 11 classifier classifier NOUN iajs-3467 77 12 is be AUX iajs-3467 77 13 : : PUNCT iajs-3467 77 14 1 1 NUM iajs-3467 77 15 ǁ𝛽ǁ̂ ǁ𝛽ǁ̂ NOUN iajs-3467 77 16 . . PUNCT iajs-3467 78 1 3 3 X iajs-3467 78 2 . . X iajs-3467 78 3 in in ADP iajs-3467 78 4 machine machine NOUN iajs-3467 78 5 learning learn VERB iajs-3467 78 6 literature literature PROPN iajs-3467 78 7 { { PUNCT iajs-3467 78 8 𝑠𝑖 𝑠𝑖 PROPN iajs-3467 78 9 } } PUNCT iajs-3467 78 10 are be AUX iajs-3467 78 11 called call VERB iajs-3467 78 12 slack slack NOUN iajs-3467 78 13 variables variable NOUN iajs-3467 78 14 . . PUNCT iajs-3467 79 1 4 4 X iajs-3467 79 2 . . X iajs-3467 79 3 in in ADP iajs-3467 79 4 this this DET iajs-3467 79 5 case case NOUN iajs-3467 79 6 , , PUNCT iajs-3467 79 7 the the DET iajs-3467 79 8 support support NOUN iajs-3467 79 9 vectors vector NOUN iajs-3467 79 10 are be AUX iajs-3467 79 11 defined define VERB iajs-3467 79 12 as as ADP iajs-3467 79 13 those those DET iajs-3467 79 14 observations observation NOUN iajs-3467 79 15 in in ADP iajs-3467 79 16 the the DET iajs-3467 79 17 training training NOUN iajs-3467 79 18 data datum NOUN iajs-3467 79 19 lies lie VERB iajs-3467 79 20 exactly exactly ADV iajs-3467 79 21 on on ADP iajs-3467 79 22 any any DET iajs-3467 79 23 one one NUM iajs-3467 79 24 of of ADP iajs-3467 79 25 the the DET iajs-3467 79 26 two two NUM iajs-3467 79 27 correct correct ADJ iajs-3467 79 28 margins margin NOUN iajs-3467 79 29 , , PUNCT iajs-3467 79 30 𝑦𝑖(Ξ²Μ‚0 𝑦𝑖(Ξ²Μ‚0 PROPN iajs-3467 79 31 + + CCONJ iajs-3467 79 32 β̂𝑇π‘₯(𝑖 β̂𝑇π‘₯(𝑖 NOUN iajs-3467 79 33 ) ) PUNCT iajs-3467 79 34 ) ) PUNCT iajs-3467 80 1 = = SYM iajs-3467 80 2 +1 +1 PROPN iajs-3467 80 3 , , PUNCT iajs-3467 80 4 or or CCONJ iajs-3467 80 5 violate violate VERB iajs-3467 80 6 the the DET iajs-3467 80 7 correct correct ADJ iajs-3467 80 8 margins margin NOUN iajs-3467 80 9 , , PUNCT iajs-3467 80 10 𝑦𝑖(Ξ²Μ‚0 𝑦𝑖(Ξ²Μ‚0 PROPN iajs-3467 80 11 + + CCONJ iajs-3467 80 12 β̂𝑇π‘₯(𝑖 β̂𝑇π‘₯(𝑖 NOUN iajs-3467 80 13 ) ) PUNCT iajs-3467 80 14 ) ) PUNCT iajs-3467 81 1 < < X iajs-3467 82 1 +1 +1 X iajs-3467 82 2 . . PUNCT iajs-3467 83 1 5 5 X iajs-3467 83 2 . . PUNCT iajs-3467 84 1 it it PRON iajs-3467 84 2 turns turn VERB iajs-3467 84 3 out out ADP iajs-3467 84 4 that that SCONJ iajs-3467 84 5 the the DET iajs-3467 84 6 estimations estimation NOUN iajs-3467 84 7 Ξ²Μ‚0 Ξ²Μ‚0 X iajs-3467 84 8 , , PUNCT iajs-3467 84 9 Ξ²Μ‚ Ξ²Μ‚ PART iajs-3467 84 10 depend depend VERB iajs-3467 84 11 only only ADV iajs-3467 84 12 on on ADP iajs-3467 84 13 the the DET iajs-3467 84 14 data data NOUN iajs-3467 84 15 points point NOUN iajs-3467 84 16 which which PRON iajs-3467 84 17 are be AUX iajs-3467 84 18 support support NOUN iajs-3467 84 19 vectors vector NOUN iajs-3467 84 20 . . PUNCT iajs-3467 85 1 that that PRON iajs-3467 85 2 means mean VERB iajs-3467 85 3 all all DET iajs-3467 85 4 data datum NOUN iajs-3467 85 5 points point NOUN iajs-3467 85 6 inside inside ADP iajs-3467 85 7 the the DET iajs-3467 85 8 correct correct ADJ iajs-3467 85 9 margins margin NOUN iajs-3467 85 10 have have VERB iajs-3467 85 11 no no DET iajs-3467 85 12 role role NOUN iajs-3467 85 13 in in ADP iajs-3467 85 14 determining determine VERB iajs-3467 85 15 the the DET iajs-3467 85 16 form form NOUN iajs-3467 85 17 of of ADP iajs-3467 85 18 the the DET iajs-3467 85 19 classifier classifier NOUN iajs-3467 85 20 . . PUNCT iajs-3467 86 1 this this DET iajs-3467 86 2 property property NOUN iajs-3467 86 3 makes make VERB iajs-3467 86 4 svm svm ADJ iajs-3467 86 5 very very ADV iajs-3467 86 6 useful useful ADJ iajs-3467 86 7 for for ADP iajs-3467 86 8 massive massive ADJ iajs-3467 86 9 data data NOUN iajs-3467 86 10 classification classification NOUN iajs-3467 86 11 problems problem NOUN iajs-3467 86 12 . . PUNCT iajs-3467 87 1 for for ADP iajs-3467 87 2 any any DET iajs-3467 87 3 two two NUM iajs-3467 87 4 nonlinearly nonlinearly ADV iajs-3467 87 5 separable separable ADJ iajs-3467 87 6 classes class NOUN iajs-3467 87 7 ( ( PUNCT iajs-3467 87 8 e.g. e.g. ADV iajs-3467 87 9 , , PUNCT iajs-3467 87 10 because because SCONJ iajs-3467 87 11 of of ADP iajs-3467 87 12 the the DET iajs-3467 87 13 noise noise NOUN iajs-3467 87 14 ) ) PUNCT iajs-3467 87 15 , , PUNCT iajs-3467 87 16 the the DET iajs-3467 87 17 optimal optimal ADJ iajs-3467 87 18 hyper hyper ADJ iajs-3467 87 19 - - ADJ iajs-3467 87 20 plane plane NOUN iajs-3467 87 21 condition condition NOUN iajs-3467 87 22 including include VERB iajs-3467 87 23 an an DET iajs-3467 87 24 extra extra ADJ iajs-3467 87 25 term term NOUN iajs-3467 87 26 can can AUX iajs-3467 87 27 be be AUX iajs-3467 87 28 formalized formalize VERB iajs-3467 87 29 as as SCONJ iajs-3467 87 30 follows follow VERB iajs-3467 87 31 : : PUNCT iajs-3467 88 1 π’šπ’Š(π‘₯𝑖 π’šπ’Š(π‘₯𝑖 PROPN iajs-3467 88 2 𝑇w 𝑇w PROPN iajs-3467 88 3 + + CCONJ iajs-3467 88 4 b b X iajs-3467 88 5 ) ) PUNCT iajs-3467 88 6 β‰₯ β‰₯ NOUN iajs-3467 88 7 1 1 NUM iajs-3467 88 8 βˆ’ βˆ’ NOUN iajs-3467 89 1   NOUN iajs-3467 89 2 𝑖 𝑖 X iajs-3467 89 3 , , PUNCT iajs-3467 89 4 , , PUNCT iajs-3467 89 5 𝑖 𝑖 NOUN iajs-3467 89 6 = = NOUN iajs-3467 89 7 1 1 NUM iajs-3467 89 8 , , PUNCT iajs-3467 89 9 . . PUNCT iajs-3467 89 10 . . PUNCT iajs-3467 89 11 . . PUNCT iajs-3467 90 1 , , PUNCT iajs-3467 90 2 𝑛. 𝑛. NOUN iajs-3467 90 3 the the DET iajs-3467 90 4 objective objective ADJ iajs-3467 90 5 function function NOUN iajs-3467 90 6 should should AUX iajs-3467 90 7 be be AUX iajs-3467 90 8 minimized minimize VERB iajs-3467 90 9 , , PUNCT iajs-3467 90 10 i.e. i.e. X iajs-3467 90 11 , , PUNCT iajs-3467 90 12   X iajs-3467 90 13 𝑖 𝑖 PRON iajs-3467 90 14 β‰₯ β‰₯ NOUN iajs-3467 90 15 0 0 NUM iajs-3467 90 16 should should AUX iajs-3467 90 17 be be AUX iajs-3467 90 18 minimized minimize VERB iajs-3467 90 19 as as ADV iajs-3467 90 20 well well ADV iajs-3467 90 21 as as ADP iajs-3467 90 22 ǁ𝑀ǁ ǁ𝑀ǁ NOUN iajs-3467 90 23 , , PUNCT iajs-3467 90 24 as as SCONJ iajs-3467 90 25 follows follow VERB iajs-3467 90 26 : : PUNCT iajs-3467 90 27 π‘šπ‘–π‘›π‘–π‘šπ‘–π‘§π‘’ π‘šπ‘–π‘›π‘–π‘šπ‘–π‘§π‘’ PROPN iajs-3467 90 28 𝑀𝑇𝑀 𝑀𝑇𝑀 PROPN iajs-3467 90 29 + + CCONJ iajs-3467 90 30 𝑐 𝑐 NOUN iajs-3467 90 31 βˆ‘ βˆ‘ PUNCT iajs-3467 90 32   NOUN iajs-3467 90 33 𝑖 𝑖 X iajs-3467 90 34 π‘˜π‘› π‘˜π‘› INTJ iajs-3467 90 35 𝑖=1 𝑖=1 VERB iajs-3467 90 36 subject subject VERB iajs-3467 90 37 to to ADP iajs-3467 90 38 𝑦𝑖(π‘₯𝑖 𝑦𝑖(π‘₯𝑖 PROPN iajs-3467 90 39 𝑇𝑀 𝑇𝑀 PROPN iajs-3467 90 40 + + CCONJ iajs-3467 90 41 𝑏 𝑏 NOUN iajs-3467 90 42 ) ) PUNCT iajs-3467 90 43 β‰₯ β‰₯ NOUN iajs-3467 90 44 1 1 NUM iajs-3467 90 45 βˆ’ βˆ’ NOUN iajs-3467 91 1   NOUN iajs-3467 91 2 𝑖 𝑖 NOUN iajs-3467 92 1 and and CCONJ iajs-3467 92 2   X iajs-3467 92 3 𝑖 𝑖 PRON iajs-3467 92 4 β‰₯ β‰₯ NOUN iajs-3467 92 5 0 0 NUM iajs-3467 92 6 ; ; PUNCT iajs-3467 92 7 𝑖 𝑖 SYM iajs-3467 92 8 = = SYM iajs-3467 92 9 1 1 NUM iajs-3467 92 10 , , PUNCT iajs-3467 92 11 . . PUNCT iajs-3467 92 12 . . PUNCT iajs-3467 92 13 . . PUNCT iajs-3467 93 1 , , PUNCT iajs-3467 93 2 𝑛. 𝑛. NOUN iajs-3467 93 3 ( ( PUNCT iajs-3467 93 4 3 3 X iajs-3467 93 5 ) ) PUNCT iajs-3467 93 6 in in ADP iajs-3467 93 7 eq eq NOUN iajs-3467 93 8 ( ( PUNCT iajs-3467 93 9 3 3 NUM iajs-3467 93 10 ) ) PUNCT iajs-3467 93 11 , , PUNCT iajs-3467 93 12 c c PROPN iajs-3467 93 13 is be AUX iajs-3467 93 14 a a DET iajs-3467 93 15 regularization regularization NOUN iajs-3467 93 16 parameter parameter NOUN iajs-3467 93 17 that that PRON iajs-3467 93 18 controls control VERB iajs-3467 93 19 the the DET iajs-3467 93 20 balance balance NOUN iajs-3467 93 21 between between ADP iajs-3467 93 22 making make VERB iajs-3467 93 23 the the DET iajs-3467 93 24 margin margin NOUN iajs-3467 93 25 as as ADV iajs-3467 93 26 big big ADJ iajs-3467 93 27 as as ADP iajs-3467 93 28 possible possible ADJ iajs-3467 93 29 and and CCONJ iajs-3467 93 30 making make VERB iajs-3467 93 31 the the DET iajs-3467 93 32 training training NOUN iajs-3467 93 33 error error NOUN iajs-3467 93 34 as as ADV iajs-3467 93 35 small small ADJ iajs-3467 93 36 as as ADP iajs-3467 93 37 possible possible ADJ iajs-3467 93 38 . . PUNCT iajs-3467 94 1 small small ADJ iajs-3467 94 2 𝐢 𝐢 PROPN iajs-3467 94 3 tends tend VERB iajs-3467 94 4 to to PART iajs-3467 94 5 focus focus VERB iajs-3467 94 6 on on ADP iajs-3467 94 7 the the DET iajs-3467 94 8 margin margin NOUN iajs-3467 94 9 and and CCONJ iajs-3467 94 10 ignore ignore VERB iajs-3467 94 11 outliers outlier NOUN iajs-3467 94 12 in in ADP iajs-3467 94 13 the the DET iajs-3467 94 14 training training NOUN iajs-3467 94 15 data datum NOUN iajs-3467 94 16 , , PUNCT iajs-3467 94 17 while while SCONJ iajs-3467 94 18 large large ADJ iajs-3467 94 19 𝐢 𝐢 PROPN iajs-3467 94 20 may may AUX iajs-3467 94 21 cause cause VERB iajs-3467 94 22 the the DET iajs-3467 94 23 training training NOUN iajs-3467 94 24 data datum NOUN iajs-3467 94 25 to to PART iajs-3467 94 26 be be AUX iajs-3467 94 27 too too ADV iajs-3467 94 28 well well ADV iajs-3467 94 29 fit fit ADJ iajs-3467 94 30 ( ( PUNCT iajs-3467 94 31 see see VERB iajs-3467 94 32 figure figure NOUN iajs-3467 94 33 3 3 NUM iajs-3467 94 34 ) ) PUNCT iajs-3467 94 35 . . PUNCT iajs-3467 95 1 3.1 3.1 NUM iajs-3467 95 2 second second ADJ iajs-3467 95 3 norm norm NOUN iajs-3467 95 4 soft soft ADJ iajs-3467 95 5 margin margin NOUN iajs-3467 95 6 the the DET iajs-3467 95 7 optimization optimization NOUN iajs-3467 95 8 problem problem NOUN iajs-3467 95 9 is be AUX iajs-3467 95 10 called call VERB iajs-3467 95 11 the the DET iajs-3467 95 12 second second ADJ iajs-3467 95 13 soft soft ADJ iajs-3467 95 14 ( ( PUNCT iajs-3467 95 15 nonlinear nonlinear ADJ iajs-3467 95 16 ) ) PUNCT iajs-3467 95 17 margin margin NOUN iajs-3467 95 18 problem problem NOUN iajs-3467 95 19 when when SCONJ iajs-3467 95 20 π‘˜ π‘˜ PROPN iajs-3467 95 21 = = SYM iajs-3467 95 22 2 2 NUM iajs-3467 95 23 , , PUNCT iajs-3467 95 24 i.e. i.e. X iajs-3467 95 25 , , PUNCT iajs-3467 95 26 π‘šπ‘–π‘›π‘–π‘šπ‘–π‘§π‘’ π‘šπ‘–π‘›π‘–π‘šπ‘–π‘§π‘’ PROPN iajs-3467 95 27 𝑀𝑇𝑀 𝑀𝑇𝑀 PROPN iajs-3467 95 28 + + CCONJ iajs-3467 95 29 𝑐 𝑐 NOUN iajs-3467 95 30 βˆ‘ βˆ‘ PUNCT iajs-3467 95 31   X iajs-3467 95 32 𝑖 𝑖 X iajs-3467 95 33 2 2 NUM iajs-3467 95 34 π‘š π‘š NOUN iajs-3467 95 35 𝑖=1 𝑖=1 PROPN iajs-3467 95 36 . . PUNCT iajs-3467 96 1 subject subject ADJ iajs-3467 96 2 to to ADP iajs-3467 96 3 𝑦𝑖(π‘₯𝑖 𝑦𝑖(π‘₯𝑖 PROPN iajs-3467 96 4 𝑇𝑀 𝑇𝑀 PROPN iajs-3467 96 5 + + CCONJ iajs-3467 96 6 𝑏 𝑏 NOUN iajs-3467 96 7 ) ) PUNCT iajs-3467 96 8 β‰₯ β‰₯ NOUN iajs-3467 96 9 1 1 NUM iajs-3467 96 10 βˆ’ βˆ’ NOUN iajs-3467 96 11   NOUN iajs-3467 96 12 𝑖 𝑖 X iajs-3467 96 13 , , PUNCT iajs-3467 96 14 ( ( PUNCT iajs-3467 96 15 𝑖 𝑖 SYM iajs-3467 96 16 = = NOUN iajs-3467 96 17 1 1 NUM iajs-3467 96 18 , , PUNCT iajs-3467 96 19 . . PUNCT iajs-3467 96 20 . . PUNCT iajs-3467 97 1 . . PUNCT iajs-3467 98 1 , , PUNCT iajs-3467 98 2 π‘š π‘š NOUN iajs-3467 98 3 ) ) PUNCT iajs-3467 98 4 . . PUNCT iajs-3467 99 1 not not PART iajs-3467 99 2 that that SCONJ iajs-3467 99 3 if if SCONJ iajs-3467 99 4   X iajs-3467 99 5 𝑖 𝑖 PRON iajs-3467 99 6 β‰₯ β‰₯ NOUN iajs-3467 99 7 0 0 NUM iajs-3467 99 8 is be AUX iajs-3467 99 9 declined decline VERB iajs-3467 99 10 , , PUNCT iajs-3467 99 11 similarly similarly ADV iajs-3467 99 12 if if SCONJ iajs-3467 99 13   PROPN iajs-3467 99 14 𝑖 𝑖 X iajs-3467 99 15 < < X iajs-3467 99 16 0 0 NUM iajs-3467 99 17 , , PUNCT iajs-3467 99 18 we we PRON iajs-3467 99 19 can can AUX iajs-3467 99 20 set set VERB iajs-3467 99 21 it it PRON iajs-3467 99 22 to to ADP iajs-3467 99 23 zero zero NUM iajs-3467 99 24 and and CCONJ iajs-3467 99 25 the the DET iajs-3467 99 26 above above ADJ iajs-3467 99 27 function function NOUN iajs-3467 99 28 can can AUX iajs-3467 99 29 be be AUX iajs-3467 99 30 further far ADV iajs-3467 99 31 reduced reduce VERB iajs-3467 99 32 . . PUNCT iajs-3467 100 1 the the DET iajs-3467 100 2 initial initial ADJ iajs-3467 100 3 lagrangian lagrangian NOUN iajs-3467 100 4 for for ADP iajs-3467 100 5 the the DET iajs-3467 100 6 above above ADJ iajs-3467 100 7 2 2 NUM iajs-3467 100 8 - - PUNCT iajs-3467 100 9 norm norm NOUN iajs-3467 100 10 problem problem NOUN iajs-3467 100 11 is be AUX iajs-3467 100 12 𝐿𝑝(𝑀 𝐿𝑝(𝑀 VERB iajs-3467 100 13 , , PUNCT iajs-3467 100 14 𝑏 𝑏 NOUN iajs-3467 100 15 , , PUNCT iajs-3467 100 16   NOUN iajs-3467 100 17 , , PUNCT iajs-3467 100 18 𝛼 𝛼 NOUN iajs-3467 100 19 ) ) PUNCT iajs-3467 100 20 = = SYM iajs-3467 100 21 1 1 NUM iajs-3467 100 22 2 2 NUM iajs-3467 100 23 𝑀𝑇𝑀 𝑀𝑇𝑀 X iajs-3467 100 24 + + NOUN iajs-3467 100 25 𝑐 𝑐 NOUN iajs-3467 100 26 2 2 NUM iajs-3467 100 27 βˆ‘ βˆ‘ NOUN iajs-3467 100 28   X iajs-3467 100 29 𝑖 𝑖 SYM iajs-3467 100 30 2 2 NUM iajs-3467 100 31 𝑛 𝑛 PRON iajs-3467 100 32 𝑖=1 𝑖=1 PUNCT iajs-3467 101 1 βˆ’ βˆ’ PROPN iajs-3467 102 1 βˆ‘ βˆ‘ PROPN iajs-3467 102 2 𝛼𝑖 𝛼𝑖 PROPN iajs-3467 102 3 𝑛 𝑛 PRON iajs-3467 102 4 𝑖=1 𝑖=1 PROPN iajs-3467 103 1 [ [ X iajs-3467 103 2 𝑦𝑖(𝑀𝑇π‘₯ 𝑦𝑖(𝑀𝑇π‘₯ PROPN iajs-3467 103 3 + + CCONJ iajs-3467 103 4 𝑏 𝑏 NOUN iajs-3467 103 5 ) ) PUNCT iajs-3467 103 6 βˆ’ βˆ’ PROPN iajs-3467 103 7 1 1 NUM iajs-3467 104 1 + + NUM iajs-3467 104 2   X iajs-3467 104 3 𝑖 𝑖 X iajs-3467 104 4 ] ] X iajs-3467 104 5 . . PUNCT iajs-3467 105 1 substituting substitute VERB iajs-3467 105 2 πœ•πΏ πœ•πΏ PRON iajs-3467 105 3 πœ•π‘€ πœ•π‘€ NOUN iajs-3467 105 4 = = PUNCT iajs-3467 105 5 w w ADP iajs-3467 105 6 βˆ’ βˆ’ NOUN iajs-3467 105 7 βˆ‘ βˆ‘ PUNCT iajs-3467 105 8 𝑦𝑖𝛼𝑖π‘₯𝑖 𝑦𝑖𝛼𝑖π‘₯𝑖 VERB iajs-3467 105 9 𝑛 𝑛 PRON iajs-3467 105 10 𝑖=1 𝑖=1 PUNCT iajs-3467 105 11 = = SYM iajs-3467 105 12 0 0 NUM iajs-3467 105 13 ; ; PUNCT iajs-3467 105 14 πœ•πΏ πœ•πΏ VERB iajs-3467 105 15 πœ•οΊ πœ•οΊ PROPN iajs-3467 105 16 = = PUNCT iajs-3467 105 17 𝑐 𝑐 PUNCT iajs-3467 105 18 βˆ’ βˆ’ NOUN iajs-3467 105 19 𝛼 𝛼 NOUN iajs-3467 105 20 = = NOUN iajs-3467 105 21 0 0 NUM iajs-3467 105 22 ; ; PUNCT iajs-3467 105 23 πœ•πΏ πœ•πΏ X iajs-3467 105 24 πœ•π‘ πœ•π‘ ADJ iajs-3467 105 25 = = SYM iajs-3467 105 26 βˆ‘ βˆ‘ PUNCT iajs-3467 105 27 𝑦𝑖𝛼𝑖 𝑦𝑖𝛼𝑖 PROPN iajs-3467 105 28 𝑛 𝑛 PRON iajs-3467 105 29 𝑖=1 𝑖=1 PUNCT iajs-3467 106 1 = = SYM iajs-3467 106 2 0 0 X iajs-3467 106 3 . . PUNCT iajs-3467 107 1 into into ADP iajs-3467 107 2 the the DET iajs-3467 107 3 initial initial ADJ iajs-3467 107 4 lagrangian lagrangian NOUN iajs-3467 107 5 , , PUNCT iajs-3467 107 6 we we PRON iajs-3467 107 7 get get VERB iajs-3467 107 8 the the DET iajs-3467 107 9 dual dual ADJ iajs-3467 107 10 problem problem NOUN iajs-3467 107 11 ihjpas ihjpa VERB iajs-3467 107 12 . . PUNCT iajs-3467 108 1 37 37 NUM iajs-3467 108 2 ( ( PUNCT iajs-3467 108 3 1 1 NUM iajs-3467 108 4 ) ) PUNCT iajs-3467 108 5 2024 2024 NUM iajs-3467 108 6 417 417 NUM iajs-3467 108 7 π‘šπ‘Žπ‘₯π‘–π‘šπ‘–π‘§π‘’ π‘šπ‘Žπ‘₯π‘–π‘šπ‘–π‘§π‘’ PROPN iajs-3467 108 8 𝐿𝑑(𝛼 𝐿𝑑(𝛼 NUM iajs-3467 108 9 ) ) PUNCT iajs-3467 108 10 = = PUNCT iajs-3467 108 11 βˆ‘ βˆ‘ PUNCT iajs-3467 108 12 𝛼𝑖 𝛼𝑖 ADP iajs-3467 108 13 π‘š π‘š X iajs-3467 108 14 𝑖=1 𝑖=1 PROPN iajs-3467 108 15 βˆ’ βˆ’ NUM iajs-3467 108 16 1 1 NUM iajs-3467 108 17 2 2 NUM iajs-3467 108 18 βˆ‘ βˆ‘ PUNCT iajs-3467 108 19 βˆ‘ βˆ‘ VERB iajs-3467 108 20 𝑦𝑖𝑦𝑗𝛼𝑖𝛼𝑗π‘₯𝑗 𝑦𝑖𝑦𝑗𝛼𝑖𝛼𝑗π‘₯𝑗 NOUN iajs-3467 108 21 𝑇π‘₯𝑖 𝑇π‘₯𝑖 PROPN iajs-3467 108 22 βˆ’ βˆ’ PROPN iajs-3467 108 23 1 1 NUM iajs-3467 108 24 2𝑐 2𝑐 NUM iajs-3467 108 25 βˆ‘ βˆ‘ PUNCT iajs-3467 108 26 𝛼𝑖 𝛼𝑖 PROPN iajs-3467 108 27 2 2 NUM iajs-3467 108 28 π‘š π‘š NOUN iajs-3467 108 29 𝑖=1 𝑖=1 PROPN iajs-3467 108 30 π‘š π‘š X iajs-3467 108 31 𝑗=1 𝑗=1 X iajs-3467 108 32 π‘š π‘š NOUN iajs-3467 108 33 𝑖=1 𝑖=1 PUNCT iajs-3467 108 34 = = PUNCT iajs-3467 108 35 βˆ‘ βˆ‘ PUNCT iajs-3467 108 36 𝛼𝑖 𝛼𝑖 ADP iajs-3467 108 37 π‘š π‘š X iajs-3467 108 38 𝑖=1 𝑖=1 PROPN iajs-3467 108 39 βˆ’ βˆ’ NUM iajs-3467 108 40 1 1 NUM iajs-3467 108 41 2 2 NUM iajs-3467 108 42 βˆ‘ βˆ‘ PROPN iajs-3467 108 43 βˆ‘ βˆ‘ PROPN iajs-3467 108 44 𝑦𝑖𝑦𝑗𝛼𝑖𝛼𝑗 𝑦𝑖𝑦𝑗𝛼𝑖𝛼𝑗 PROPN iajs-3467 108 45 π‘š π‘š PROPN iajs-3467 108 46 𝑗=1 𝑗=1 PROPN iajs-3467 108 47 π‘š π‘š PROPN iajs-3467 108 48 𝑖=1 𝑖=1 PROPN iajs-3467 108 49 ( ( PUNCT iajs-3467 108 50 π‘₯𝑗 π‘₯𝑗 NOUN iajs-3467 108 51 𝑇π‘₯𝑖 𝑇π‘₯𝑖 PROPN iajs-3467 108 52 + + CCONJ iajs-3467 108 53 1 1 NUM iajs-3467 108 54 𝑐 𝑐 PROPN iajs-3467 108 55 𝛿𝑖𝑗 𝛿𝑖𝑗 NOUN iajs-3467 108 56 ) ) PUNCT iajs-3467 108 57 . . PUNCT iajs-3467 109 1 subject subject ADJ iajs-3467 109 2 to to ADP iajs-3467 109 3 𝛼𝑖 𝛼𝑖 PROPN iajs-3467 109 4 β‰₯ β‰₯ NUM iajs-3467 109 5 0 0 NUM iajs-3467 109 6 , , PUNCT iajs-3467 109 7 βˆ‘ βˆ‘ PROPN iajs-3467 109 8 𝛼𝑖 𝛼𝑖 ADP iajs-3467 109 9 π‘š π‘š X iajs-3467 109 10 𝑖=1 𝑖=1 PUNCT iajs-3467 109 11 𝑦𝑖 𝑦𝑖 PROPN iajs-3467 109 12 = = NOUN iajs-3467 109 13 0 0 PROPN iajs-3467 109 14 . . PUNCT iajs-3467 110 1 the the DET iajs-3467 110 2 above above ADJ iajs-3467 110 3 quadratic quadratic ADJ iajs-3467 110 4 programming programming NOUN iajs-3467 110 5 can can AUX iajs-3467 110 6 be be AUX iajs-3467 110 7 solved solve VERB iajs-3467 110 8 for for ADP iajs-3467 110 9 𝛼𝑖. 𝛼𝑖. PROPN iajs-3467 110 10 as as ADP iajs-3467 110 11 a a DET iajs-3467 110 12 result result NOUN iajs-3467 110 13 , , PUNCT iajs-3467 110 14 all all DET iajs-3467 110 15 support support NOUN iajs-3467 110 16 vectors vector NOUN iajs-3467 110 17 that that PRON iajs-3467 110 18 correspond correspond VERB iajs-3467 110 19 to to ADP iajs-3467 110 20 π‘₯𝑖 π‘₯𝑖 PRON iajs-3467 110 21 > > X iajs-3467 110 22 0 0 PUNCT iajs-3467 111 1 satisfy satisfy VERB iajs-3467 111 2 the the DET iajs-3467 111 3 following follow VERB iajs-3467 111 4 condition condition NOUN iajs-3467 111 5 [ [ X iajs-3467 111 6 8 8 NUM iajs-3467 111 7 ] ] X iajs-3467 111 8 : : PUNCT iajs-3467 112 1 𝑦𝑖(𝑋𝑖 𝑦𝑖(𝑋𝑖 PROPN iajs-3467 112 2 𝑇𝑀 𝑇𝑀 PROPN iajs-3467 112 3 + + CCONJ iajs-3467 112 4 𝑏 𝑏 NOUN iajs-3467 112 5 ) ) PUNCT iajs-3467 112 6 = = SYM iajs-3467 113 1 1 1 NUM iajs-3467 113 2 βˆ’ βˆ’ NOUN iajs-3467 113 3 πœ‰π‘– πœ‰π‘– X iajs-3467 113 4 . . PUNCT iajs-3467 114 1 substituting substitute VERB iajs-3467 114 2 𝑀 𝑀 ADP iajs-3467 114 3 = = PUNCT iajs-3467 114 4 βˆ‘ βˆ‘ PUNCT iajs-3467 114 5 𝑦𝑗𝑦𝑗π‘₯π‘—π‘—πœ–π‘  𝑦𝑗𝑦𝑗π‘₯π‘—π‘—πœ–π‘  NOUN iajs-3467 114 6 into into ADP iajs-3467 114 7 this this DET iajs-3467 114 8 equation equation NOUN iajs-3467 114 9 ( ( PUNCT iajs-3467 114 10 where where SCONJ iajs-3467 114 11 s s NOUN iajs-3467 114 12 is be AUX iajs-3467 114 13 the the DET iajs-3467 114 14 set set NOUN iajs-3467 114 15 of of ADP iajs-3467 114 16 support support NOUN iajs-3467 114 17 vector vector NOUN iajs-3467 114 18 ) ) PUNCT iajs-3467 114 19 , , PUNCT iajs-3467 114 20 we we PRON iajs-3467 114 21 get get VERB iajs-3467 114 22 : : PUNCT iajs-3467 114 23 𝑦𝑖(βˆ‘ 𝑦𝑖(βˆ‘ NOUN iajs-3467 114 24 𝑦𝑗𝛼𝑗(𝑋𝑖 𝑦𝑗𝛼𝑗(𝑋𝑖 NOUN iajs-3467 114 25 𝑇𝑋𝑗 𝑇𝑋𝑗 PROPN iajs-3467 114 26 ) ) PUNCT iajs-3467 115 1 + + CCONJ iajs-3467 115 2 π‘π‘—βˆˆπ‘† π‘π‘—βˆˆπ‘† X iajs-3467 115 3 ) ) PUNCT iajs-3467 116 1 = = SYM iajs-3467 116 2 1 1 NUM iajs-3467 116 3 βˆ’ βˆ’ PROPN iajs-3467 117 1 πœ‰π‘– πœ‰π‘– PROPN iajs-3467 117 2 , , PUNCT iajs-3467 117 3 𝑖. 𝑖. ADV iajs-3467 117 4 𝑒. 𝑒. PROPN iajs-3467 117 5 , , PUNCT iajs-3467 117 6 𝑦𝑖 𝑦𝑖 PROPN iajs-3467 117 7 βˆ‘ βˆ‘ PROPN iajs-3467 117 8 π‘¦π‘—π›Όπ‘—π‘—βˆˆπ‘† π‘¦π‘—π›Όπ‘—π‘—βˆˆπ‘† PROPN iajs-3467 117 9 ( ( PUNCT iajs-3467 117 10 π‘₯𝑖 π‘₯𝑖 X iajs-3467 117 11 𝑇π‘₯𝑗 𝑇π‘₯𝑗 PROPN iajs-3467 117 12 ) ) PUNCT iajs-3467 117 13 = = SYM iajs-3467 118 1 1 1 NUM iajs-3467 118 2 βˆ’ βˆ’ NOUN iajs-3467 118 3 πœ‰π‘– πœ‰π‘– INTJ iajs-3467 118 4 βˆ’ βˆ’ PROPN iajs-3467 118 5 𝑦𝑖𝑏 𝑦𝑖𝑏 PROPN iajs-3467 118 6 . . PUNCT iajs-3467 119 1 for for ADP iajs-3467 119 2 the the DET iajs-3467 119 3 optimal optimal ADJ iajs-3467 119 4 weight weight NOUN iajs-3467 119 5 w w NOUN iajs-3467 119 6 , , PUNCT iajs-3467 119 7 we we PRON iajs-3467 119 8 have have AUX iajs-3467 119 9 : : PUNCT iajs-3467 119 10 ǁ𝑀ǁ2 ǁ𝑀ǁ2 PROPN iajs-3467 119 11 ∢ ∢ PROPN iajs-3467 119 12 𝑀𝑇𝑀 𝑀𝑇𝑀 X iajs-3467 119 13 = = PUNCT iajs-3467 119 14 βˆ‘ βˆ‘ PART iajs-3467 119 15 𝛼𝑖𝑦𝑖π‘₯𝑖 𝛼𝑖𝑦𝑖π‘₯𝑖 NOUN iajs-3467 119 16 𝑇 𝑇 PROPN iajs-3467 119 17 π‘–βˆˆπ‘† π‘–βˆˆπ‘† PRON iajs-3467 119 18 βˆ‘ βˆ‘ PUNCT iajs-3467 119 19 𝛼𝑗𝑦𝑗π‘₯𝑗 𝛼𝑗𝑦𝑗π‘₯𝑗 NOUN iajs-3467 119 20 π‘—βˆˆπ‘† π‘—βˆˆπ‘† PROPN iajs-3467 120 1 = = PUNCT iajs-3467 120 2 βˆ‘ βˆ‘ PUNCT iajs-3467 120 3 𝛼𝑖𝑦𝑖 𝛼𝑖𝑦𝑖 X iajs-3467 120 4 βˆ‘ βˆ‘ PUNCT iajs-3467 120 5 𝛼𝑖𝑦𝑖π‘₯𝑖 𝛼𝑖𝑦𝑖π‘₯𝑖 VERB iajs-3467 120 6 𝑇π‘₯𝑗 𝑇π‘₯𝑗 PROPN iajs-3467 120 7 π‘—βˆˆπ‘†π‘–βˆˆπ‘† π‘—βˆˆπ‘†π‘–βˆˆπ‘† PROPN iajs-3467 120 8 . . PUNCT iajs-3467 121 1 βˆ‘ βˆ‘ X iajs-3467 122 1 𝛼𝑖(1 𝛼𝑖(1 VERB iajs-3467 122 2 βˆ’ βˆ’ NOUN iajs-3467 122 3 πœ‰π‘– πœ‰π‘– ADV iajs-3467 122 4 βˆ’ βˆ’ NOUN iajs-3467 122 5 𝑦𝑖 𝑦𝑖 PROPN iajs-3467 122 6 𝑏 𝑏 NOUN iajs-3467 122 7 ) ) PUNCT iajs-3467 123 1 = = PUNCT iajs-3467 123 2 βˆ‘ βˆ‘ PUNCT iajs-3467 123 3 𝛼𝑖 𝛼𝑖 INTJ iajs-3467 123 4 βˆ’ βˆ’ NOUN iajs-3467 123 5 βˆ‘ βˆ‘ PUNCT iajs-3467 123 6 π›Όπ‘–πœ‰π‘– π›Όπ‘–πœ‰π‘– VERB iajs-3467 123 7 βˆ’ βˆ’ NOUN iajs-3467 123 8 𝑏 𝑏 PROPN iajs-3467 123 9 βˆ‘ βˆ‘ ADV iajs-3467 123 10 π‘¦π‘–π›Όπ‘–π‘–βˆˆπ‘†π‘–βˆˆπ‘†π‘–βˆˆπ‘†π‘–βˆˆπ‘† π‘¦π‘–π›Όπ‘–π‘–βˆˆπ‘†π‘–βˆˆπ‘†π‘–βˆˆπ‘†π‘–βˆˆπ‘† PROPN iajs-3467 123 11 . . PUNCT iajs-3467 124 1 since since SCONJ iajs-3467 124 2 πœ‰π‘– πœ‰π‘– PROPN iajs-3467 124 3 = = SYM iajs-3467 124 4 𝛼𝑖/𝐢 𝛼𝑖/𝐢 PROPN iajs-3467 124 5 , , PUNCT iajs-3467 124 6 we we PRON iajs-3467 124 7 have have AUX iajs-3467 124 8 : : PUNCT iajs-3467 124 9 βˆ‘ βˆ‘ VERB iajs-3467 124 10 π›Όπ‘–π‘–βˆˆπ‘† π›Όπ‘–π‘–βˆˆπ‘† ADJ iajs-3467 124 11 βˆ’ βˆ’ PROPN iajs-3467 124 12 βˆ‘ βˆ‘ ADV iajs-3467 124 13 π›Όπ‘–πœ‰π‘– π›Όπ‘–πœ‰π‘– ADJ iajs-3467 124 14 = = PUNCT iajs-3467 124 15 βˆ‘ βˆ‘ PUNCT iajs-3467 124 16 𝛼𝑖 𝛼𝑖 INTJ iajs-3467 124 17 βˆ’ βˆ’ NOUN iajs-3467 124 18 1 1 NUM iajs-3467 124 19 𝑐 𝑐 PROPN iajs-3467 124 20 βˆ‘ βˆ‘ PUNCT iajs-3467 124 21 𝛼𝑖 𝛼𝑖 PROPN iajs-3467 124 22 2 2 NUM iajs-3467 124 23 π‘–βˆˆπ‘†π‘–βˆˆπ‘†π‘–βˆˆπ‘† π‘–βˆˆπ‘†π‘–βˆˆπ‘†π‘–βˆˆπ‘† PUNCT iajs-3467 124 24 . . PUNCT iajs-3467 125 1 therefor therefor ADP iajs-3467 125 2 the the DET iajs-3467 125 3 optimal optimal ADJ iajs-3467 125 4 separation separation NOUN iajs-3467 125 5 margin margin NOUN iajs-3467 125 6 becomes becomes AUX iajs-3467 125 7 : : PUNCT iajs-3467 125 8 1 1 NUM iajs-3467 125 9 ǁwǁ ǁwǁ NOUN iajs-3467 125 10 = = SYM iajs-3467 125 11 ( ( PUNCT iajs-3467 125 12 βˆ‘ βˆ‘ PROPN iajs-3467 125 13 𝛼𝑖 𝛼𝑖 INTJ iajs-3467 125 14 βˆ’ βˆ’ PROPN iajs-3467 125 15 1 1 NUM iajs-3467 125 16 𝑐 𝑐 PROPN iajs-3467 125 17 βˆ‘ βˆ‘ PUNCT iajs-3467 125 18 𝛼𝑖 𝛼𝑖 PROPN iajs-3467 125 19 2 2 NUM iajs-3467 125 20 π‘–βˆˆπ‘†π‘–βˆˆπ‘† π‘–βˆˆπ‘†π‘–βˆˆπ‘† ADJ iajs-3467 125 21 ) ) PUNCT iajs-3467 125 22 βˆ’1/2 βˆ’1/2 ADJ iajs-3467 125 23 . . PUNCT iajs-3467 126 1 3.2 3.2 NUM iajs-3467 126 2 first first ADJ iajs-3467 126 3 norm norm NOUN iajs-3467 126 4 soft soft ADJ iajs-3467 126 5 margin margin NOUN iajs-3467 126 6 the the DET iajs-3467 126 7 optimization optimization NOUN iajs-3467 126 8 problem problem NOUN iajs-3467 126 9 is be AUX iajs-3467 126 10 called call VERB iajs-3467 126 11 the the DET iajs-3467 126 12 first first ADJ iajs-3467 126 13 soft soft ADJ iajs-3467 126 14 ( ( PUNCT iajs-3467 126 15 nonlinear nonlinear ADJ iajs-3467 126 16 ) ) PUNCT iajs-3467 126 17 margin margin NOUN iajs-3467 126 18 problem problem NOUN iajs-3467 126 19 when when SCONJ iajs-3467 126 20 k k PROPN iajs-3467 126 21 = = NOUN iajs-3467 126 22 1 1 NUM iajs-3467 126 23 , , PUNCT iajs-3467 126 24 i.e. i.e. X iajs-3467 126 25 , , PUNCT iajs-3467 126 26 minimize minimize VERB iajs-3467 126 27 𝑀𝑇𝑀 𝑀𝑇𝑀 SYM iajs-3467 126 28 + + NOUN iajs-3467 126 29 𝑐 𝑐 NOUN iajs-3467 126 30 βˆ‘ βˆ‘ PUNCT iajs-3467 126 31   NOUN iajs-3467 126 32 𝑖 𝑖 SYM iajs-3467 126 33 π‘š π‘š X iajs-3467 126 34 𝑖=1 𝑖=1 PUNCT iajs-3467 126 35 subject subject ADJ iajs-3467 126 36 to to ADP iajs-3467 126 37 𝑦𝑖(π‘₯𝑖 𝑦𝑖(π‘₯𝑖 ADJ iajs-3467 126 38 𝑇w 𝑇w PROPN iajs-3467 126 39 + + CCONJ iajs-3467 126 40 b b X iajs-3467 126 41 ) ) PUNCT iajs-3467 126 42 β‰₯ β‰₯ NOUN iajs-3467 126 43 1 1 NUM iajs-3467 126 44 βˆ’ βˆ’ NOUN iajs-3467 127 1   NOUN iajs-3467 127 2 𝑖 𝑖 INTJ iajs-3467 127 3 , , PUNCT iajs-3467 127 4   X iajs-3467 127 5 𝑖 𝑖 PRON iajs-3467 127 6 β‰₯ β‰₯ NOUN iajs-3467 127 7 0 0 NUM iajs-3467 127 8 ; ; PUNCT iajs-3467 127 9 i=1 i=1 X iajs-3467 127 10 , , PUNCT iajs-3467 127 11 … … PUNCT iajs-3467 127 12 , , PUNCT iajs-3467 127 13 n. n. VERB iajs-3467 127 14 the the DET iajs-3467 127 15 first first ADJ iajs-3467 127 16 norm norm NOUN iajs-3467 127 17 algorithm algorithm NOUN iajs-3467 127 18 is be AUX iajs-3467 127 19 less less ADV iajs-3467 127 20 complex complex ADJ iajs-3467 127 21 compared compare VERB iajs-3467 127 22 with with ADP iajs-3467 127 23 the the DET iajs-3467 127 24 second second ADJ iajs-3467 127 25 norm norm NOUN iajs-3467 127 26 algorithm algorithm NOUN iajs-3467 127 27 . . PUNCT iajs-3467 128 1 it it PRON iajs-3467 128 2 is be AUX iajs-3467 128 3 powerful powerful ADJ iajs-3467 128 4 when when SCONJ iajs-3467 128 5 the the DET iajs-3467 128 6 training training NOUN iajs-3467 128 7 dataset dataset NOUN iajs-3467 128 8 has have VERB iajs-3467 128 9 outliers outlier NOUN iajs-3467 128 10 . . PUNCT iajs-3467 129 1 the the DET iajs-3467 129 2 first first ADJ iajs-3467 129 3 norm norm NOUN iajs-3467 129 4 method method NOUN iajs-3467 129 5 should should AUX iajs-3467 129 6 be be AUX iajs-3467 129 7 used use VERB iajs-3467 129 8 to to PART iajs-3467 129 9 ignore ignore VERB iajs-3467 129 10 the the DET iajs-3467 129 11 outliers outlier NOUN iajs-3467 129 12 when when SCONJ iajs-3467 129 13 the the DET iajs-3467 129 14 data data NOUN iajs-3467 129 15 is be AUX iajs-3467 129 16 noisy noisy ADJ iajs-3467 129 17 [ [ X iajs-3467 129 18 9,10 9,10 NUM iajs-3467 129 19 ] ] X iajs-3467 129 20 . . PUNCT iajs-3467 130 1 the the DET iajs-3467 130 2 primal primal ADJ iajs-3467 130 3 lagrangian lagrangian NOUN iajs-3467 130 4 for for ADP iajs-3467 130 5 first first ADJ iajs-3467 130 6 norm norm NOUN iajs-3467 130 7 problem problem NOUN iajs-3467 130 8 above above ADV iajs-3467 130 9 is be AUX iajs-3467 130 10 : : PUNCT iajs-3467 130 11 𝐿𝑝(𝑀 𝐿𝑝(𝑀 ADJ iajs-3467 130 12 , , PUNCT iajs-3467 130 13 𝑏 𝑏 NOUN iajs-3467 130 14 , , PUNCT iajs-3467 130 15 πœ‰ πœ‰ X iajs-3467 130 16 , , PUNCT iajs-3467 130 17 𝛼 𝛼 X iajs-3467 130 18 , , PUNCT iajs-3467 130 19 Ο“ Ο“ NOUN iajs-3467 130 20 ) ) PUNCT iajs-3467 130 21 = = SYM iajs-3467 130 22 1 1 NUM iajs-3467 130 23 2 2 NUM iajs-3467 130 24 𝑀𝑇𝑀 𝑀𝑇𝑀 X iajs-3467 130 25 + + CCONJ iajs-3467 130 26 𝑐 𝑐 NOUN iajs-3467 130 27 βˆ‘ βˆ‘ PUNCT iajs-3467 130 28 Ο“π‘–πœ‰π‘– Ο“π‘–πœ‰π‘– PROPN iajs-3467 130 29 𝑛 𝑛 PRON iajs-3467 130 30 𝑖=1 𝑖=1 PUNCT iajs-3467 130 31 βˆ’ βˆ’ PROPN iajs-3467 131 1 βˆ‘[𝑦𝑖(𝑀𝑇π‘₯ βˆ‘[𝑦𝑖(𝑀𝑇π‘₯ ADJ iajs-3467 131 2 + + NUM iajs-3467 131 3 𝑏 𝑏 NOUN iajs-3467 131 4 ) ) PUNCT iajs-3467 131 5 βˆ’ βˆ’ PROPN iajs-3467 131 6 1 1 NUM iajs-3467 132 1 + + CCONJ iajs-3467 132 2 πœ‰π‘– πœ‰π‘– X iajs-3467 132 3 ] ] X iajs-3467 132 4 𝑛 𝑛 PRON iajs-3467 132 5 𝑖=1 𝑖=1 PROPN iajs-3467 132 6 βˆ’ βˆ’ PUNCT iajs-3467 132 7 βˆ‘ βˆ‘ PUNCT iajs-3467 133 1 Ο“π‘–πœ‰π‘– Ο“π‘–πœ‰π‘– PROPN iajs-3467 133 2 𝑛 𝑛 PRON iajs-3467 133 3 𝑖=1 𝑖=1 PROPN iajs-3467 133 4 . . PUNCT iajs-3467 134 1 with with ADP iajs-3467 134 2 𝛼𝑖 𝛼𝑖 ADP iajs-3467 134 3 β‰₯ β‰₯ X iajs-3467 134 4 0 0 NUM iajs-3467 134 5 π‘Žπ‘›π‘‘ π‘Žπ‘›π‘‘ ADJ iajs-3467 134 6 ϓ𝑖 ϓ𝑖 PROPN iajs-3467 134 7 β‰₯ β‰₯ NOUN iajs-3467 134 8 0 0 NUM iajs-3467 134 9 . . PUNCT iajs-3467 135 1 substituting substitute VERB iajs-3467 135 2 ihjpas ihjpas PROPN iajs-3467 135 3 . . PUNCT iajs-3467 136 1 37 37 NUM iajs-3467 136 2 ( ( PUNCT iajs-3467 136 3 1 1 NUM iajs-3467 136 4 ) ) PUNCT iajs-3467 136 5 2024 2024 NUM iajs-3467 136 6 418 418 NUM iajs-3467 136 7 πœ•πΏ πœ•πΏ VERB iajs-3467 136 8 πœ•π‘€ πœ•π‘€ NOUN iajs-3467 136 9 = = SYM iajs-3467 136 10 𝑀 𝑀 PART iajs-3467 136 11 βˆ’ βˆ’ NOUN iajs-3467 136 12 βˆ‘ βˆ‘ PUNCT iajs-3467 136 13 𝑦𝑖𝛼𝑖π‘₯𝑖 𝑦𝑖𝛼𝑖π‘₯𝑖 VERB iajs-3467 136 14 𝑛 𝑛 PRON iajs-3467 136 15 𝑖=1 𝑖=1 PUNCT iajs-3467 136 16 = = SYM iajs-3467 136 17 0 0 NUM iajs-3467 136 18 ; ; PUNCT iajs-3467 136 19 πœ•πΏ πœ•πΏ X iajs-3467 136 20 πœ•πœ‰ πœ•πœ‰ NOUN iajs-3467 136 21 = = SYM iajs-3467 136 22 π‘πœ‰ π‘πœ‰ PROPN iajs-3467 137 1 βˆ’ βˆ’ PROPN iajs-3467 137 2 𝛼 𝛼 NOUN iajs-3467 137 3 = = NOUN iajs-3467 137 4 0 0 NUM iajs-3467 137 5 ; ; PUNCT iajs-3467 137 6 πœ•πΏ πœ•πΏ X iajs-3467 137 7 πœ•π‘ πœ•π‘ ADJ iajs-3467 137 8 = = SYM iajs-3467 137 9 βˆ‘ βˆ‘ PUNCT iajs-3467 137 10 𝑦𝑖𝛼𝑖 𝑦𝑖𝛼𝑖 PROPN iajs-3467 137 11 = = SYM iajs-3467 137 12 0 0 PUNCT iajs-3467 137 13 𝑛 𝑛 PRON iajs-3467 137 14 𝑖=1 𝑖=1 PROPN iajs-3467 137 15 . . PUNCT iajs-3467 138 1 into into ADP iajs-3467 138 2 the the DET iajs-3467 138 3 initial initial ADJ iajs-3467 138 4 lagrangian lagrangian NOUN iajs-3467 138 5 , , PUNCT iajs-3467 138 6 we we PRON iajs-3467 138 7 get get VERB iajs-3467 138 8 the the DET iajs-3467 138 9 dual dual ADJ iajs-3467 138 10 problem problem NOUN iajs-3467 138 11 : : PUNCT iajs-3467 138 12 maximize maximize VERB iajs-3467 138 13 𝐿𝑑(𝛼 𝐿𝑑(𝛼 PROPN iajs-3467 138 14 , , PUNCT iajs-3467 138 15 Ο“ Ο“ PROPN iajs-3467 138 16 ) ) PUNCT iajs-3467 138 17 = = PUNCT iajs-3467 138 18 βˆ‘ βˆ‘ PUNCT iajs-3467 138 19 𝛼𝑖 𝛼𝑖 INTJ iajs-3467 138 20 𝑛 𝑛 PRON iajs-3467 138 21 𝑖=1 𝑖=1 PUNCT iajs-3467 139 1 βˆ’ βˆ’ NOUN iajs-3467 139 2 1 1 NUM iajs-3467 139 3 2 2 NUM iajs-3467 139 4 βˆ‘ βˆ‘ PUNCT iajs-3467 139 5 βˆ‘ βˆ‘ VERB iajs-3467 139 6 𝑦𝑖𝑦𝑗𝛼𝑖𝛼𝑗π‘₯𝑗 𝑦𝑖𝑦𝑗𝛼𝑖𝛼𝑗π‘₯𝑗 PROPN iajs-3467 139 7 𝑇π‘₯𝑖 𝑇π‘₯𝑖 PROPN iajs-3467 139 8 βˆ’ βˆ’ PROPN iajs-3467 139 9 βˆ‘ βˆ‘ PUNCT iajs-3467 139 10 π›Όπ‘–πœ‰π‘– π›Όπ‘–πœ‰π‘– VERB iajs-3467 139 11 βˆ’ βˆ’ NOUN iajs-3467 139 12 βˆ‘ βˆ‘ PUNCT iajs-3467 139 13 Ο“π‘–πœ‰π‘– Ο“π‘–πœ‰π‘– PROPN iajs-3467 139 14 + + PROPN iajs-3467 139 15 𝑛 𝑛 PROPN iajs-3467 139 16 𝑖=1 𝑖=1 NOUN iajs-3467 139 17 𝑛 𝑛 PRON iajs-3467 139 18 𝑖=1 𝑖=1 PUNCT iajs-3467 139 19 𝑛 𝑛 PRON iajs-3467 139 20 𝑗=1 𝑗=1 PROPN iajs-3467 139 21 𝑛 𝑛 PRON iajs-3467 140 1 𝑖=1 𝑖=1 PUNCT iajs-3467 140 2 𝑐 𝑐 NOUN iajs-3467 140 3 βˆ‘ βˆ‘ PUNCT iajs-3467 140 4 πœ‰π‘– πœ‰π‘– PROPN iajs-3467 140 5 = = PUNCT iajs-3467 140 6 βˆ‘ βˆ‘ PROPN iajs-3467 140 7 𝛼𝑖 𝛼𝑖 INTJ iajs-3467 140 8 𝑛 𝑛 PRON iajs-3467 140 9 𝑖=1 𝑖=1 PUNCT iajs-3467 141 1 βˆ’ βˆ’ NOUN iajs-3467 141 2 1 1 NUM iajs-3467 141 3 2 2 NUM iajs-3467 141 4 βˆ‘ βˆ‘ PUNCT iajs-3467 141 5 βˆ‘ βˆ‘ VERB iajs-3467 141 6 𝑦𝑖𝑦𝑗𝛼𝑖𝛼𝑗π‘₯𝑗 𝑦𝑖𝑦𝑗𝛼𝑖𝛼𝑗π‘₯𝑗 PROPN iajs-3467 141 7 𝑇π‘₯𝑖 𝑇π‘₯𝑖 PROPN iajs-3467 141 8 𝑛 𝑛 PRON iajs-3467 141 9 𝑖=1 𝑖=1 PROPN iajs-3467 141 10 𝑛 𝑛 PRON iajs-3467 141 11 𝑖=1 𝑖=1 PUNCT iajs-3467 141 12 𝑛 𝑛 PRON iajs-3467 141 13 𝑖=1 𝑖=1 PROPN iajs-3467 141 14 . . PUNCT iajs-3467 142 1 subject subject ADJ iajs-3467 142 2 to to ADP iajs-3467 142 3 0 0 NUM iajs-3467 142 4 ≀ ≀ NUM iajs-3467 142 5 𝛼𝑖 𝛼𝑖 PRON iajs-3467 142 6 ≀ ≀ NUM iajs-3467 142 7 𝑐 𝑐 NOUN iajs-3467 142 8 , , PUNCT iajs-3467 142 9 βˆ‘ βˆ‘ PUNCT iajs-3467 142 10 𝛼𝑖𝑦𝑖 𝛼𝑖𝑦𝑖 ADJ iajs-3467 142 11 𝑛 𝑛 PRON iajs-3467 142 12 𝑖=1 𝑖=1 PUNCT iajs-3467 143 1 = = SYM iajs-3467 143 2 0 0 X iajs-3467 143 3 . . PUNCT iajs-3467 144 1 remarkably remarkably ADV iajs-3467 144 2 the the DET iajs-3467 144 3 dual dual ADJ iajs-3467 144 4 problem problem NOUN iajs-3467 144 5 objective objective ADJ iajs-3467 144 6 function function NOUN iajs-3467 144 7 is be AUX iajs-3467 144 8 the the DET iajs-3467 144 9 same same ADJ iajs-3467 144 10 as as ADP iajs-3467 144 11 that that PRON iajs-3467 144 12 of of ADP iajs-3467 144 13 the the DET iajs-3467 144 14 linearly linearly ADV iajs-3467 144 15 separable separable ADJ iajs-3467 144 16 case case NOUN iajs-3467 144 17 that that PRON iajs-3467 144 18 was be AUX iajs-3467 144 19 discussed discuss VERB iajs-3467 144 20 previously previously ADV iajs-3467 144 21 . . PUNCT iajs-3467 145 1 this this PRON iajs-3467 145 2 is be AUX iajs-3467 145 3 because because SCONJ iajs-3467 145 4 the the DET iajs-3467 145 5 cancellation cancellation NOUN iajs-3467 145 6 depends depend VERB iajs-3467 145 7 on on ADP iajs-3467 145 8 𝑐 𝑐 PROPN iajs-3467 145 9 = = SYM iajs-3467 145 10 𝛼𝑖 𝛼𝑖 PROPN iajs-3467 145 11 + + CCONJ iajs-3467 145 12 𝑦𝑖. 𝑦𝑖. PROPN iajs-3467 145 13 now now ADV iajs-3467 145 14 , , PUNCT iajs-3467 145 15 when when SCONJ iajs-3467 145 16 𝛼𝑖 𝛼𝑖 PROPN iajs-3467 145 17 β‰₯ β‰₯ VERB iajs-3467 145 18 0 0 PUNCT iajs-3467 145 19 and and CCONJ iajs-3467 145 20 ϓ𝑖 ϓ𝑖 PROPN iajs-3467 145 21 β‰₯ β‰₯ NOUN iajs-3467 145 22 0 0 NUM iajs-3467 145 23 , , PUNCT iajs-3467 145 24 we we PRON iajs-3467 145 25 get get VERB iajs-3467 145 26 0 0 NUM iajs-3467 145 27 ≀ ≀ NUM iajs-3467 145 28 𝛼𝑖 𝛼𝑖 ADP iajs-3467 145 29 ≀ ≀ NOUN iajs-3467 145 30 𝑐. 𝑐. ADV iajs-3467 145 31 when when SCONJ iajs-3467 145 32 we we PRON iajs-3467 145 33 solve solve VERB iajs-3467 145 34 the the DET iajs-3467 145 35 quadratic quadratic ADJ iajs-3467 145 36 programming programming NOUN iajs-3467 145 37 ( ( PUNCT iajs-3467 145 38 qp qp NOUN iajs-3467 145 39 ) ) PUNCT iajs-3467 145 40 problem problem NOUN iajs-3467 145 41 for for ADP iajs-3467 145 42 𝛼𝑖 𝛼𝑖 PROPN iajs-3467 145 43 , , PUNCT iajs-3467 145 44 the the DET iajs-3467 145 45 following follow VERB iajs-3467 145 46 optimal optimal ADJ iajs-3467 145 47 decision decision NOUN iajs-3467 145 48 plane plane NOUN iajs-3467 145 49 with with ADP iajs-3467 145 50 the the DET iajs-3467 145 51 margin margin NOUN iajs-3467 145 52 can can AUX iajs-3467 145 53 be be AUX iajs-3467 145 54 gotten get VERB iajs-3467 145 55 : : PUNCT iajs-3467 145 56 ( ( PUNCT iajs-3467 145 57 βˆ‘ βˆ‘ INTJ iajs-3467 145 58 βˆ‘ βˆ‘ NUM iajs-3467 145 59 𝛼𝑖𝛼𝑗𝑦𝑖𝑦𝑗π‘₯𝑖 𝛼𝑖𝛼𝑗𝑦𝑖𝑦𝑗π‘₯𝑖 NOUN iajs-3467 145 60 𝑇π‘₯𝑗 𝑇π‘₯𝑗 PROPN iajs-3467 145 61 π‘—βˆˆπ‘†π‘–βˆˆπ‘† π‘—βˆˆπ‘†π‘–βˆˆπ‘† PROPN iajs-3467 145 62 ) ) PUNCT iajs-3467 145 63 βˆ’1/2 βˆ’1/2 VERB iajs-3467 145 64 4 4 NUM iajs-3467 145 65 . . PUNCT iajs-3467 145 66 kernels kernel NOUN iajs-3467 145 67 transformation transformation NOUN iajs-3467 145 68 the the DET iajs-3467 145 69 kernel kernel PROPN iajs-3467 145 70 transformation transformation NOUN iajs-3467 145 71 techniques technique NOUN iajs-3467 145 72 can can AUX iajs-3467 145 73 be be AUX iajs-3467 145 74 applied apply VERB iajs-3467 145 75 when when SCONJ iajs-3467 145 76 , , PUNCT iajs-3467 145 77 the the DET iajs-3467 145 78 condition condition NOUN iajs-3467 145 79 set set NOUN iajs-3467 145 80 of of ADP iajs-3467 145 81 karush karush NOUN iajs-3467 145 82 - - PUNCT iajs-3467 145 83 kuhntucker kuhntucker PROPN iajs-3467 145 84 ( ( PUNCT iajs-3467 145 85 kkt kkt PROPN iajs-3467 145 86 ) ) PUNCT iajs-3467 145 87 are be AUX iajs-3467 145 88 satisfied satisfied ADJ iajs-3467 145 89 . . PUNCT iajs-3467 146 1 eq eq X iajs-3467 146 2 ( ( PUNCT iajs-3467 146 3 1 1 X iajs-3467 146 4 ) ) PUNCT iajs-3467 146 5 is be AUX iajs-3467 146 6 linear linear ADJ iajs-3467 146 7 in in ADP iajs-3467 146 8 terms term NOUN iajs-3467 146 9 of of ADP iajs-3467 146 10 the the DET iajs-3467 146 11 new new ADJ iajs-3467 146 12 space space NOUN iajs-3467 146 13 that that PRON iajs-3467 146 14 ΙΈ ΙΈ X iajs-3467 146 15 ( ( PUNCT iajs-3467 146 16 x x X iajs-3467 146 17 ) ) PUNCT iajs-3467 146 18 maps map VERB iajs-3467 146 19 the the DET iajs-3467 146 20 data datum NOUN iajs-3467 146 21 to to ADP iajs-3467 146 22 non non ADJ iajs-3467 146 23 - - ADJ iajs-3467 146 24 linear linear ADJ iajs-3467 146 25 in in ADP iajs-3467 146 26 the the DET iajs-3467 146 27 space space NOUN iajs-3467 146 28 , , PUNCT iajs-3467 146 29 see see VERB iajs-3467 146 30 figure figure NOUN iajs-3467 146 31 3 3 NUM iajs-3467 146 32 . . PUNCT iajs-3467 147 1 the the DET iajs-3467 147 2 most most ADV iajs-3467 147 3 common common ADJ iajs-3467 147 4 kernels kernel NOUN iajs-3467 147 5 are be AUX iajs-3467 147 6 : : PUNCT iajs-3467 147 7 linear linear ADJ iajs-3467 147 8 , , PUNCT iajs-3467 147 9 polynomial polynomial ADJ iajs-3467 147 10 , , PUNCT iajs-3467 147 11 sigmoid sigmoid NOUN iajs-3467 147 12 or or CCONJ iajs-3467 147 13 multi multi ADJ iajs-3467 147 14 - - ADJ iajs-3467 147 15 layer layer ADJ iajs-3467 147 16 perceptron perceptron NOUN iajs-3467 147 17 ( ( PUNCT iajs-3467 147 18 mlp mlp PROPN iajs-3467 147 19 ) ) PUNCT iajs-3467 147 20 and and CCONJ iajs-3467 147 21 gaussian gaussian ADJ iajs-3467 147 22 or or CCONJ iajs-3467 147 23 radial radial ADJ iajs-3467 147 24 basis basis NOUN iajs-3467 147 25 function function NOUN iajs-3467 147 26 ( ( PUNCT iajs-3467 147 27 rpf rpf NOUN iajs-3467 147 28 ) ) PUNCT iajs-3467 148 1 [ [ X iajs-3467 148 2 11 11 NUM iajs-3467 148 3 - - SYM iajs-3467 148 4 13 13 NUM iajs-3467 148 5 ] ] PUNCT iajs-3467 148 6 . . PUNCT iajs-3467 149 1 their their PRON iajs-3467 149 2 expressions expression NOUN iajs-3467 149 3 are be AUX iajs-3467 149 4 as as SCONJ iajs-3467 149 5 follows follow VERB iajs-3467 149 6 : : PUNCT iajs-3467 149 7 linear linear PROPN iajs-3467 149 8 kernel:𝐾(π‘₯𝑖 kernel:𝐾(π‘₯𝑖 PROPN iajs-3467 149 9 , , PUNCT iajs-3467 149 10 π‘₯𝑗 π‘₯𝑗 PROPN iajs-3467 149 11 ) ) PUNCT iajs-3467 149 12 = = SYM iajs-3467 150 1 π‘₯𝑖 π‘₯𝑖 X iajs-3467 150 2 𝑇π‘₯𝑗 𝑇π‘₯𝑗 PROPN iajs-3467 150 3 . . PUNCT iajs-3467 151 1 polynomial polynomial PROPN iajs-3467 151 2 kernel:𝐾(π‘₯𝑖 kernel:𝐾(π‘₯𝑖 PROPN iajs-3467 151 3 , , PUNCT iajs-3467 151 4 π‘₯𝑗 π‘₯𝑗 PROPN iajs-3467 151 5 ) ) PUNCT iajs-3467 151 6 = = PUNCT iajs-3467 151 7 ( ( PUNCT iajs-3467 151 8 1 1 NUM iajs-3467 151 9 + + NUM iajs-3467 151 10 π‘₯𝑖 π‘₯𝑖 PROPN iajs-3467 151 11 𝑇π‘₯𝑗)𝑝. 𝑇π‘₯𝑗)𝑝. PROPN iajs-3467 151 12 sigmoid sigmoid NOUN iajs-3467 151 13 ( ( PUNCT iajs-3467 151 14 mlp mlp PROPN iajs-3467 151 15 ) ) PUNCT iajs-3467 151 16 kernel kernel PROPN iajs-3467 151 17 : : PUNCT iajs-3467 151 18 𝐾(π‘₯𝑖 𝐾(π‘₯𝑖 ADJ iajs-3467 151 19 , , PUNCT iajs-3467 151 20 π‘₯𝑗 π‘₯𝑗 PROPN iajs-3467 151 21 ) ) PUNCT iajs-3467 152 1 = = SYM iajs-3467 152 2 π‘‘π‘Žπ‘›β„Ž(π‘˜1π‘₯𝑖 π‘‘π‘Žπ‘›β„Ž(π‘˜1π‘₯𝑖 PROPN iajs-3467 152 3 𝑇π‘₯𝑗 𝑇π‘₯𝑗 PROPN iajs-3467 152 4 + + CCONJ iajs-3467 152 5 π‘˜2 π‘˜2 PROPN iajs-3467 152 6 ) ) PUNCT iajs-3467 152 7 . . PUNCT iajs-3467 153 1 gaussian gaussian NOUN iajs-3467 153 2 ( ( PUNCT iajs-3467 153 3 rbf rbf PROPN iajs-3467 153 4 ) ) PUNCT iajs-3467 153 5 kernel kernel PROPN iajs-3467 153 6 : : PUNCT iajs-3467 153 7 𝐾(π‘₯𝑖 𝐾(π‘₯𝑖 ADJ iajs-3467 153 8 , , PUNCT iajs-3467 153 9 π‘₯𝑗 π‘₯𝑗 PROPN iajs-3467 153 10 ) ) PUNCT iajs-3467 153 11 = = PUNCT iajs-3467 153 12 𝑒π‘₯𝑝 𝑒π‘₯𝑝 PROPN iajs-3467 153 13 [ [ PUNCT iajs-3467 153 14 βˆ’(π‘₯π‘–βˆ’π‘₯𝑗)𝑇(π‘₯π‘–βˆ’π‘₯𝑗 βˆ’(π‘₯π‘–βˆ’π‘₯𝑗)𝑇(π‘₯π‘–βˆ’π‘₯𝑗 NOUN iajs-3467 153 15 ) ) PUNCT iajs-3467 153 16 2𝜎2 2𝜎2 NUM iajs-3467 153 17 ] ] PUNCT iajs-3467 153 18 figure figure NOUN iajs-3467 153 19 3 3 NUM iajs-3467 153 20 . . PUNCT iajs-3467 154 1 tow tow NOUN iajs-3467 154 2 - - PUNCT iajs-3467 154 3 dimensional dimensional ADJ iajs-3467 154 4 space space NOUN iajs-3467 154 5 verses verses ADP iajs-3467 154 6 three three NUM iajs-3467 154 7 - - PUNCT iajs-3467 154 8 dimensional dimensional ADJ iajs-3467 154 9 space space NOUN iajs-3467 154 10 . . PUNCT iajs-3467 155 1 ihjpas ihjpas PROPN iajs-3467 155 2 . . PUNCT iajs-3467 156 1 37 37 NUM iajs-3467 156 2 ( ( PUNCT iajs-3467 156 3 1 1 NUM iajs-3467 156 4 ) ) PUNCT iajs-3467 156 5 2024 2024 NUM iajs-3467 156 6 419 419 NUM iajs-3467 156 7 we we PRON iajs-3467 156 8 define define VERB iajs-3467 156 9 the the DET iajs-3467 156 10 kernel kernel NOUN iajs-3467 156 11 function function NOUN iajs-3467 156 12 as as ADP iajs-3467 156 13 k(π‘₯𝑖 k(π‘₯𝑖 PROPN iajs-3467 156 14 , , PUNCT iajs-3467 156 15 π‘₯𝑗 π‘₯𝑗 PROPN iajs-3467 156 16 ) ) PUNCT iajs-3467 157 1 = = NOUN iajs-3467 157 2 < < X iajs-3467 157 3 ΙΈ(π‘₯𝑖 ΙΈ(π‘₯𝑖 PROPN iajs-3467 157 4 ) ) PUNCT iajs-3467 157 5 , , PUNCT iajs-3467 157 6 ΙΈ(π‘₯𝑗 ΙΈ(π‘₯𝑗 PROPN iajs-3467 157 7 ) ) PUNCT iajs-3467 157 8 > > PUNCT iajs-3467 157 9 = = SYM iajs-3467 157 10 ΙΈ(π‘₯𝑖)𝑇ɸ(π‘₯𝑗 ΙΈ(π‘₯𝑖)𝑇ɸ(π‘₯𝑗 NOUN iajs-3467 157 11 ) ) PUNCT iajs-3467 157 12 where where SCONJ iajs-3467 157 13 ΙΈ ΙΈ NOUN iajs-3467 157 14 is be AUX iajs-3467 157 15 a a DET iajs-3467 157 16 mapping mapping NOUN iajs-3467 157 17 from from ADP iajs-3467 157 18 input input NOUN iajs-3467 157 19 space space NOUN iajs-3467 157 20 to to ADP iajs-3467 157 21 output output NOUN iajs-3467 157 22 space space NOUN iajs-3467 157 23 , , PUNCT iajs-3467 157 24 see see VERB iajs-3467 157 25 figure figure NOUN iajs-3467 157 26 4 4 NUM iajs-3467 157 27 . . PUNCT iajs-3467 158 1 ( ( PUNCT iajs-3467 158 2 a a X iajs-3467 158 3 ) ) PUNCT iajs-3467 158 4 with with ADP iajs-3467 158 5 xΜ… xΜ… PROPN iajs-3467 158 6 = = SYM iajs-3467 158 7 ( ( PUNCT iajs-3467 158 8 x1 x1 PROPN iajs-3467 158 9 , , PUNCT iajs-3467 158 10 x2 x2 ADJ iajs-3467 158 11 ) ) PUNCT iajs-3467 158 12 ∈ ∈ PROPN iajs-3467 158 13 r2 r2 PROPN iajs-3467 158 14 ito ito PROPN iajs-3467 158 15 a a DET iajs-3467 158 16 feature feature NOUN iajs-3467 158 17 space space NOUN iajs-3467 158 18 ( ( PUNCT iajs-3467 158 19 b b NOUN iajs-3467 158 20 ) ) PUNCT iajs-3467 158 21 with with ADP iajs-3467 158 22 𝑧̅ 𝑧̅ NOUN iajs-3467 158 23 = = SYM iajs-3467 158 24 ( ( PUNCT iajs-3467 158 25 𝑧1 𝑧1 PROPN iajs-3467 158 26 , , PUNCT iajs-3467 158 27 𝑧2 𝑧2 NOUN iajs-3467 158 28 , , PUNCT iajs-3467 158 29 𝑧3 𝑧3 PROPN iajs-3467 158 30 ) ) PUNCT iajs-3467 158 31 ∈ ∈ PROPN iajs-3467 158 32 r3 r3 PROPN iajs-3467 158 33 , , PUNCT iajs-3467 158 34 𝑒𝑠𝑖𝑛𝑔 𝑒𝑠𝑖𝑛𝑔 ADJ iajs-3467 158 35   PROPN iajs-3467 158 36 : : PUNCT iajs-3467 158 37 𝑅2 𝑅2 PROPN iajs-3467 158 38 β†’ β†’ SYM iajs-3467 158 39 𝑅3 𝑅3 PROPN iajs-3467 158 40 figure figure NOUN iajs-3467 158 41 4 4 NUM iajs-3467 158 42 . . PUNCT iajs-3467 158 43 transforming transform VERB iajs-3467 158 44 a a DET iajs-3467 158 45 nonlinear nonlinear ADJ iajs-3467 158 46 data datum NOUN iajs-3467 158 47 set set VERB iajs-3467 158 48 using use VERB iajs-3467 158 49 a a DET iajs-3467 158 50 kernel kernel NOUN iajs-3467 158 51 . . PUNCT iajs-3467 159 1 now now ADV iajs-3467 159 2 , , PUNCT iajs-3467 159 3 the the DET iajs-3467 159 4 corresponding corresponding ADJ iajs-3467 159 5 dual dual ADJ iajs-3467 159 6 form form NOUN iajs-3467 159 7 is be AUX iajs-3467 159 8 𝐿(𝛼 𝐿(𝛼 PRON iajs-3467 159 9 ) ) PUNCT iajs-3467 159 10 = = PUNCT iajs-3467 160 1 βˆ‘ βˆ‘ PUNCT iajs-3467 160 2 𝛼𝑖 𝛼𝑖 INTJ iajs-3467 160 3 𝑛 𝑛 PRON iajs-3467 160 4 𝑖=1 𝑖=1 PUNCT iajs-3467 161 1 βˆ’ βˆ’ NOUN iajs-3467 161 2 1 1 NUM iajs-3467 161 3 2 2 NUM iajs-3467 161 4 βˆ‘ βˆ‘ DET iajs-3467 161 5 βˆ‘ βˆ‘ ADV iajs-3467 161 6 𝛼𝑖𝛼𝑗𝑦𝑖𝑦𝑗𝐾(π‘₯𝑖 𝛼𝑖𝛼𝑗𝑦𝑖𝑦𝑗𝐾(π‘₯𝑖 ADJ iajs-3467 161 7 , , PUNCT iajs-3467 161 8 π‘₯𝑗)𝑛 π‘₯𝑗)𝑛 PROPN iajs-3467 161 9 𝑗=1 𝑗=1 PROPN iajs-3467 161 10 𝑛 𝑛 PROPN iajs-3467 161 11 𝑖=1 𝑖=1 PROPN iajs-3467 161 12 . . PUNCT iajs-3467 162 1 subject subject ADJ iajs-3467 162 2 to to ADP iajs-3467 162 3 βˆ‘ βˆ‘ PUNCT iajs-3467 162 4 𝛼𝑖𝑦𝑖 𝛼𝑖𝑦𝑖 PROPN iajs-3467 162 5 𝑛 𝑛 PRON iajs-3467 162 6 𝑖=1 𝑖=1 PUNCT iajs-3467 162 7 = = SYM iajs-3467 162 8 0 0 NUM iajs-3467 162 9 𝛼𝑖 𝛼𝑖 NUM iajs-3467 162 10 β‰₯ β‰₯ NOUN iajs-3467 162 11 0 0 NUM iajs-3467 162 12 , , PUNCT iajs-3467 162 13 𝑖 𝑖 NOUN iajs-3467 162 14 = = NOUN iajs-3467 162 15 1 1 NUM iajs-3467 162 16 , , PUNCT iajs-3467 162 17 . . PUNCT iajs-3467 162 18 . . PUNCT iajs-3467 162 19 . . PUNCT iajs-3467 163 1 , , PUNCT iajs-3467 163 2 𝑛. 𝑛. NOUN iajs-3467 163 3 ( ( PUNCT iajs-3467 163 4 4 4 NUM iajs-3467 163 5 ) ) PUNCT iajs-3467 163 6 eq eq NOUN iajs-3467 163 7 ( ( PUNCT iajs-3467 163 8 4 4 X iajs-3467 163 9 ) ) PUNCT iajs-3467 163 10 is be AUX iajs-3467 163 11 called call VERB iajs-3467 163 12 a a DET iajs-3467 163 13 cost cost NOUN iajs-3467 163 14 function function NOUN iajs-3467 163 15 , , PUNCT iajs-3467 163 16 and and CCONJ iajs-3467 163 17 it it PRON iajs-3467 163 18 is be AUX iajs-3467 163 19 convex convex ADJ iajs-3467 163 20 and and CCONJ iajs-3467 163 21 quadratic quadratic ADJ iajs-3467 163 22 in in ADP iajs-3467 163 23 terms term NOUN iajs-3467 163 24 of of ADP iajs-3467 163 25 the the DET iajs-3467 163 26 unknown unknown ADJ iajs-3467 163 27 parameters parameter NOUN iajs-3467 163 28 . . PUNCT iajs-3467 164 1 it it PRON iajs-3467 164 2 can can AUX iajs-3467 164 3 be be AUX iajs-3467 164 4 solved solve VERB iajs-3467 164 5 using use VERB iajs-3467 164 6 quadratic quadratic ADJ iajs-3467 164 7 programming programming NOUN iajs-3467 164 8 . . PUNCT iajs-3467 165 1 the the DET iajs-3467 165 2 final final ADJ iajs-3467 165 3 decision decision NOUN iajs-3467 165 4 rule rule NOUN iajs-3467 165 5 for for ADP iajs-3467 165 6 classification classification NOUN iajs-3467 165 7 using use VERB iajs-3467 165 8 kkt kkt PROPN iajs-3467 165 9 conditions condition NOUN iajs-3467 165 10 is be AUX iajs-3467 165 11 [ [ X iajs-3467 165 12 14,15 14,15 NUM iajs-3467 165 13 ] ] PUNCT iajs-3467 165 14 : : PUNCT iajs-3467 165 15 𝐿(π‘₯ 𝐿(π‘₯ PROPN iajs-3467 165 16 , , PUNCT iajs-3467 165 17 π›Όβˆ— π›Όβˆ— PROPN iajs-3467 165 18 , , PUNCT iajs-3467 165 19 𝛽0 𝛽0 PROPN iajs-3467 165 20 ) ) PUNCT iajs-3467 165 21 = = PUNCT iajs-3467 166 1 βˆ‘ βˆ‘ PUNCT iajs-3467 166 2 𝑦𝑖𝛼𝑖 𝑦𝑖𝛼𝑖 PROPN iajs-3467 166 3 βˆ— βˆ— PROPN iajs-3467 166 4 𝑁𝑆 𝑁𝑆 PROPN iajs-3467 166 5 𝑖=1 𝑖=1 PROPN iajs-3467 166 6 𝐾(π‘₯𝑖 𝐾(π‘₯𝑖 ADJ iajs-3467 166 7 , , PUNCT iajs-3467 166 8 π‘₯ π‘₯ NOUN iajs-3467 166 9 ) ) PUNCT iajs-3467 166 10 + + CCONJ iajs-3467 167 1 𝛽0 𝛽0 ADJ iajs-3467 167 2 . . PUNCT iajs-3467 168 1 where where SCONJ iajs-3467 168 2 𝑁𝑆 𝑁𝑆 PROPN iajs-3467 168 3 is be AUX iajs-3467 168 4 the the DET iajs-3467 168 5 number number NOUN iajs-3467 168 6 of of ADP iajs-3467 168 7 support support NOUN iajs-3467 168 8 vectors vector NOUN iajs-3467 168 9 , , PUNCT iajs-3467 168 10 and and CCONJ iajs-3467 168 11 𝛼𝑖 𝛼𝑖 ADP iajs-3467 168 12 the the DET iajs-3467 168 13 non non ADJ iajs-3467 168 14 - - ADJ iajs-3467 168 15 zero zero NUM iajs-3467 168 16 lagrange lagrange NOUN iajs-3467 168 17 multipliers multiplier NOUN iajs-3467 168 18 that that PRON iajs-3467 168 19 associated associate VERB iajs-3467 168 20 with with ADP iajs-3467 168 21 the the DET iajs-3467 168 22 support support NOUN iajs-3467 168 23 vectors vector NOUN iajs-3467 168 24 . . PUNCT iajs-3467 169 1 5 5 X iajs-3467 169 2 . . NUM iajs-3467 169 3 enhanced enhance VERB iajs-3467 169 4 stochastic stochastic ADJ iajs-3467 169 5 gradient gradient ADJ iajs-3467 169 6 descent descent NOUN iajs-3467 169 7 svm svm VERB iajs-3467 169 8 the the DET iajs-3467 169 9 goal goal NOUN iajs-3467 169 10 of of ADP iajs-3467 169 11 this this DET iajs-3467 169 12 section section NOUN iajs-3467 169 13 is be AUX iajs-3467 169 14 to to PART iajs-3467 169 15 minimize minimize VERB iajs-3467 169 16 the the DET iajs-3467 169 17 following follow VERB iajs-3467 169 18 function function NOUN iajs-3467 169 19 , , PUNCT iajs-3467 169 20 𝐿(𝛽 𝐿(𝛽 NUM iajs-3467 169 21 ) ) PUNCT iajs-3467 169 22 = = SYM iajs-3467 169 23 1 1 NUM iajs-3467 169 24 2 2 NUM iajs-3467 169 25 𝛽0 𝛽0 NOUN iajs-3467 169 26 𝑇𝛽0 𝑇𝛽0 NOUN iajs-3467 170 1 + + CCONJ iajs-3467 170 2 𝐾 𝐾 PROPN iajs-3467 170 3 βˆ‘ βˆ‘ PROPN iajs-3467 170 4 max max PROPN iajs-3467 170 5 ( ( PUNCT iajs-3467 170 6 0 0 NUM iajs-3467 170 7 , , PUNCT iajs-3467 170 8 1 1 NUM iajs-3467 170 9 βˆ’ βˆ’ NOUN iajs-3467 170 10 𝑦𝑖𝛽 𝑦𝑖𝛽 NOUN iajs-3467 171 1 𝑇π‘₯𝑖 𝑇π‘₯𝑖 PROPN iajs-3467 171 2 ) ) PUNCT iajs-3467 171 3 𝑖 𝑖 SYM iajs-3467 171 4 ( ( PUNCT iajs-3467 171 5 5 5 NUM iajs-3467 171 6 ) ) PUNCT iajs-3467 171 7 equation(5 equation(5 NUM iajs-3467 171 8 ) ) PUNCT iajs-3467 171 9 is be AUX iajs-3467 171 10 a a DET iajs-3467 171 11 quadratic quadratic ADJ iajs-3467 171 12 optimization optimization NOUN iajs-3467 171 13 problem problem NOUN iajs-3467 171 14 and and CCONJ iajs-3467 171 15 it it PRON iajs-3467 171 16 is be AUX iajs-3467 171 17 convex convex ADJ iajs-3467 171 18 in in ADP iajs-3467 171 19 𝑝. 𝑝. NOUN iajs-3467 171 20 in in ADP iajs-3467 171 21 the the DET iajs-3467 171 22 previous previous ADJ iajs-3467 171 23 section section NOUN iajs-3467 171 24 , , PUNCT iajs-3467 171 25 the the DET iajs-3467 171 26 qp qp NOUN iajs-3467 171 27 technique technique NOUN iajs-3467 171 28 was be AUX iajs-3467 171 29 used use VERB iajs-3467 171 30 , , PUNCT iajs-3467 171 31 but but CCONJ iajs-3467 171 32 it it PRON iajs-3467 171 33 is be AUX iajs-3467 171 34 very very ADV iajs-3467 171 35 slow slow ADJ iajs-3467 171 36 . . PUNCT iajs-3467 172 1 if if SCONJ iajs-3467 172 2 there there PRON iajs-3467 172 3 are be VERB iajs-3467 172 4 no no DET iajs-3467 172 5 constraints constraint NOUN iajs-3467 172 6 , , PUNCT iajs-3467 172 7 gradient gradient ADJ iajs-3467 172 8 descent descent NOUN iajs-3467 172 9 can can AUX iajs-3467 172 10 be be AUX iajs-3467 172 11 used use VERB iajs-3467 172 12 [ [ PUNCT iajs-3467 172 13 16,17 16,17 NUM iajs-3467 172 14 ] ] PUNCT iajs-3467 172 15 . . PUNCT iajs-3467 173 1 in in ADP iajs-3467 173 2 general general ADJ iajs-3467 173 3 , , PUNCT iajs-3467 173 4 the the DET iajs-3467 173 5 gradient gradient NOUN iajs-3467 173 6 goes go VERB iajs-3467 173 7 in in ADP iajs-3467 173 8 the the DET iajs-3467 173 9 opposite opposite ADJ iajs-3467 173 10 direction direction NOUN iajs-3467 173 11 to to PART iajs-3467 173 12 get get VERB iajs-3467 173 13 to to ADP iajs-3467 173 14 the the DET iajs-3467 173 15 minimum minimum NOUN iajs-3467 173 16 , , PUNCT iajs-3467 173 17 as as SCONJ iajs-3467 173 18 shown show VERB iajs-3467 173 19 in in ADP iajs-3467 173 20 figure figure NOUN iajs-3467 173 21 5.a 5.a NUM iajs-3467 173 22 , , PUNCT iajs-3467 173 23 because because SCONJ iajs-3467 173 24 the the DET iajs-3467 173 25 function function NOUN iajs-3467 173 26 is be AUX iajs-3467 173 27 in in ADP iajs-3467 173 28 the the DET iajs-3467 173 29 direction direction NOUN iajs-3467 173 30 of of ADP iajs-3467 173 31 the the DET iajs-3467 173 32 steepest steep ADJ iajs-3467 173 33 slope slope NOUN iajs-3467 173 34 . . PUNCT iajs-3467 174 1 ihjpas ihjpas PROPN iajs-3467 174 2 . . PUNCT iajs-3467 175 1 37 37 NUM iajs-3467 175 2 ( ( PUNCT iajs-3467 175 3 1 1 NUM iajs-3467 175 4 ) ) PUNCT iajs-3467 175 5 2024 2024 NUM iajs-3467 175 6 420 420 NUM iajs-3467 175 7 a. a. NOUN iajs-3467 175 8 minimizing minimizing NOUN iajs-3467 175 9 𝐿(𝛽 𝐿(𝛽 NUM iajs-3467 175 10 ) ) PUNCT iajs-3467 175 11 using use VERB iajs-3467 175 12 a a DET iajs-3467 175 13 general general ADJ iajs-3467 175 14 strategy strategy NOUN iajs-3467 175 15 b. b. PROPN iajs-3467 175 16 minimizing minimize VERB iajs-3467 175 17 𝐿(𝛽 𝐿(𝛽 NUM iajs-3467 175 18 ) ) PUNCT iajs-3467 175 19 using use VERB iajs-3467 175 20 sgd sgd NOUN iajs-3467 175 21 figure figure NOUN iajs-3467 175 22 5.minimizing 5.minimizing PROPN iajs-3467 175 23 𝐿(𝛽 𝐿(𝛽 NOUN iajs-3467 175 24 ) ) PUNCT iajs-3467 175 25 using use VERB iajs-3467 175 26 stochastic stochastic ADJ iajs-3467 175 27 gradient gradient ADJ iajs-3467 175 28 descent descent NOUN iajs-3467 175 29 . . PUNCT iajs-3467 176 1 the the DET iajs-3467 176 2 general general ADJ iajs-3467 176 3 gradient gradient ADJ iajs-3467 176 4 descent descent NOUN iajs-3467 176 5 svm svm NOUN iajs-3467 176 6 strategy strategy NOUN iajs-3467 176 7 ( ( PUNCT iajs-3467 176 8 gd gd NOUN iajs-3467 176 9 - - PUNCT iajs-3467 176 10 svm svm NOUN iajs-3467 176 11 ) ) PUNCT iajs-3467 176 12 for for ADP iajs-3467 176 13 minimizing minimize VERB iajs-3467 176 14 eq.(5 eq.(5 NOUN iajs-3467 176 15 ) ) PUNCT iajs-3467 176 16 starts start VERB iajs-3467 176 17 with with ADP iajs-3467 176 18 an an DET iajs-3467 176 19 initial initial ADJ iajs-3467 176 20 value value NOUN iajs-3467 176 21 for for ADP iajs-3467 176 22 𝛽 𝛽 NOUN iajs-3467 176 23 , , PUNCT iajs-3467 176 24 say say VERB iajs-3467 176 25 𝛽0 𝛽0 PROPN iajs-3467 176 26 , , PUNCT iajs-3467 176 27 then then ADV iajs-3467 176 28 iterate iterate NOUN iajs-3467 176 29 until until ADP iajs-3467 176 30 convergence convergence NOUN iajs-3467 176 31 . . PUNCT iajs-3467 177 1 gd gd ADJ iajs-3467 177 2 - - PUNCT iajs-3467 177 3 svm svm PROPN iajs-3467 177 4 is be AUX iajs-3467 177 5 faster fast ADJ iajs-3467 177 6 than than ADP iajs-3467 177 7 qp qp NOUN iajs-3467 177 8 , , PUNCT iajs-3467 177 9 but but CCONJ iajs-3467 177 10 it it PRON iajs-3467 177 11 is be AUX iajs-3467 177 12 still still ADV iajs-3467 177 13 slow slow ADJ iajs-3467 177 14 because because SCONJ iajs-3467 177 15 computing compute VERB iajs-3467 177 16 βˆ‡(𝛽𝑗 βˆ‡(𝛽𝑗 PROPN iajs-3467 177 17 ) ) PUNCT iajs-3467 177 18 takes take VERB iajs-3467 177 19 𝑝(𝑝 𝑝(𝑝 PROPN iajs-3467 177 20 ) ) PUNCT iajs-3467 177 21 time time NOUN iajs-3467 177 22 in in ADP iajs-3467 177 23 complexity complexity NOUN iajs-3467 177 24 , , PUNCT iajs-3467 177 25 where where SCONJ iajs-3467 177 26 𝑝 𝑝 NOUN iajs-3467 177 27 is be AUX iajs-3467 177 28 the the DET iajs-3467 177 29 size size NOUN iajs-3467 177 30 of of ADP iajs-3467 177 31 the the DET iajs-3467 177 32 training training NOUN iajs-3467 177 33 dataset dataset NOUN iajs-3467 177 34 . . PUNCT iajs-3467 178 1 if if SCONJ iajs-3467 178 2 𝑝 𝑝 PROPN iajs-3467 178 3 is be AUX iajs-3467 178 4 large large ADJ iajs-3467 178 5 , , PUNCT iajs-3467 178 6 gd gd ADJ iajs-3467 178 7 - - PUNCT iajs-3467 178 8 svm svm NOUN iajs-3467 178 9 is be AUX iajs-3467 178 10 slow slow ADJ iajs-3467 178 11 [ [ X iajs-3467 178 12 18,19 18,19 NUM iajs-3467 178 13 ] ] PUNCT iajs-3467 178 14 . . PUNCT iajs-3467 179 1 in in ADP iajs-3467 179 2 stochastic stochastic ADJ iajs-3467 179 3 gradient gradient ADJ iajs-3467 179 4 descent descent NOUN iajs-3467 179 5 ( ( PUNCT iajs-3467 179 6 sgd sgd PROPN iajs-3467 179 7 ) ) PUNCT iajs-3467 179 8 , , PUNCT iajs-3467 179 9 the the DET iajs-3467 179 10 value value NOUN iajs-3467 179 11 of of ADP iajs-3467 179 12 the the DET iajs-3467 179 13 objective objective ADJ iajs-3467 179 14 function function NOUN iajs-3467 179 15 is be AUX iajs-3467 179 16 improved improve VERB iajs-3467 179 17 at at ADP iajs-3467 179 18 each each DET iajs-3467 179 19 step step NOUN iajs-3467 179 20 . . PUNCT iajs-3467 180 1 evaluating evaluate VERB iajs-3467 180 2 the the DET iajs-3467 180 3 gradient gradient NOUN iajs-3467 180 4 for for ADP iajs-3467 180 5 each each DET iajs-3467 180 6 training training NOUN iajs-3467 180 7 sample sample NOUN iajs-3467 180 8 instead instead ADV iajs-3467 180 9 of of ADP iajs-3467 180 10 evaluating evaluate VERB iajs-3467 180 11 it it PRON iajs-3467 180 12 for for ADP iajs-3467 180 13 all all DET iajs-3467 180 14 samples sample NOUN iajs-3467 180 15 speeds speed NOUN iajs-3467 180 16 up up ADP iajs-3467 180 17 the the DET iajs-3467 180 18 process process NOUN iajs-3467 180 19 . . PUNCT iajs-3467 181 1 this this DET iajs-3467 181 2 pressure pressure NOUN iajs-3467 181 3 is be AUX iajs-3467 181 4 called call VERB iajs-3467 181 5 enhanced enhanced ADJ iajs-3467 181 6 stochastic stochastic ADJ iajs-3467 181 7 gradient gradient ADJ iajs-3467 181 8 descent descent NOUN iajs-3467 181 9 svm svm NOUN iajs-3467 181 10 ( ( PUNCT iajs-3467 181 11 esgd esgd NOUN iajs-3467 181 12 - - PUNCT iajs-3467 181 13 svm svm NOUN iajs-3467 181 14 ) ) PUNCT iajs-3467 181 15 . . PUNCT iajs-3467 182 1 as as SCONJ iajs-3467 182 2 can can AUX iajs-3467 182 3 be be AUX iajs-3467 182 4 seen see VERB iajs-3467 182 5 in in ADP iajs-3467 182 6 fig.5.b fig.5.b NOUN iajs-3467 182 7 , , PUNCT iajs-3467 182 8 esgd esgd NOUN iajs-3467 182 9 - - PUNCT iajs-3467 182 10 svm svm PROPN iajs-3467 182 11 requires require VERB iajs-3467 182 12 many many ADJ iajs-3467 182 13 more more ADJ iajs-3467 182 14 steps step NOUN iajs-3467 182 15 than than ADP iajs-3467 182 16 the the DET iajs-3467 182 17 gd gd NOUN iajs-3467 182 18 - - PUNCT iajs-3467 182 19 svm svm NOUN iajs-3467 182 20 method method NOUN iajs-3467 182 21 , , PUNCT iajs-3467 182 22 but but CCONJ iajs-3467 182 23 it it PRON iajs-3467 182 24 is be AUX iajs-3467 182 25 less less ADV iajs-3467 182 26 computationally computationally ADV iajs-3467 182 27 intensive intensive ADJ iajs-3467 182 28 at at ADP iajs-3467 182 29 each each DET iajs-3467 182 30 update update NOUN iajs-3467 182 31 ; ; PUNCT iajs-3467 182 32 and and CCONJ iajs-3467 182 33 esgd esgd NOUN iajs-3467 182 34 - - PUNCT iajs-3467 182 35 svm svm PROPN iajs-3467 182 36 is be AUX iajs-3467 182 37 faster fast ADJ iajs-3467 182 38 than than ADP iajs-3467 182 39 the the DET iajs-3467 182 40 gd gd NOUN iajs-3467 182 41 - - PUNCT iajs-3467 182 42 svm svm NOUN iajs-3467 182 43 method method NOUN iajs-3467 182 44 . . PUNCT iajs-3467 183 1 the the DET iajs-3467 183 2 esgd esgd NOUN iajs-3467 183 3 - - PUNCT iajs-3467 183 4 svm svm NOUN iajs-3467 183 5 algorithm algorithm NOUN iajs-3467 183 6 ( ( PUNCT iajs-3467 183 7 algorithm algorithm NOUN iajs-3467 183 8 1 1 NUM iajs-3467 183 9 ) ) PUNCT iajs-3467 183 10 is be AUX iajs-3467 183 11 guaranteed guarantee VERB iajs-3467 183 12 to to PART iajs-3467 183 13 converge converge VERB iajs-3467 183 14 to to ADP iajs-3467 183 15 the the DET iajs-3467 183 16 minimum minimum NOUN iajs-3467 183 17 of of ADP iajs-3467 183 18 𝑝 𝑝 NOUN iajs-3467 183 19 if if SCONJ iajs-3467 183 20 , , PUNCT iajs-3467 183 21 𝑝-𝑝. 𝑝-𝑝. VERB iajs-3467 183 22 is be AUX iajs-3467 183 23 small small ADJ iajs-3467 183 24 enough enough ADV iajs-3467 183 25 [ [ X iajs-3467 183 26 20 20 NUM iajs-3467 183 27 - - SYM iajs-3467 183 28 22 22 NUM iajs-3467 183 29 ] ] PUNCT iajs-3467 183 30 . . PUNCT iajs-3467 184 1 algorithm algorithm PROPN iajs-3467 184 2 1 1 NUM iajs-3467 184 3 : : PUNCT iajs-3467 184 4 svm svm VERB iajs-3467 184 5 using use VERB iajs-3467 184 6 stochastic stochastic ADJ iajs-3467 184 7 gradient gradient ADJ iajs-3467 184 8 descent descent NOUN iajs-3467 184 9 ( ( PUNCT iajs-3467 184 10 esgd esgd NOUN iajs-3467 184 11 - - PUNCT iajs-3467 184 12 svm svm NOUN iajs-3467 184 13 ) ) PUNCT iajs-3467 184 14 given give VERB iajs-3467 184 15 a a DET iajs-3467 184 16 training training NOUN iajs-3467 184 17 set set VERB iajs-3467 184 18 𝑆 𝑆 PROPN iajs-3467 184 19 = = SYM iajs-3467 184 20 { { PUNCT iajs-3467 184 21 ( ( PUNCT iajs-3467 184 22 π‘₯𝑖 π‘₯𝑖 PROPN iajs-3467 184 23 , , PUNCT iajs-3467 184 24 𝑦𝑖 𝑦𝑖 PROPN iajs-3467 184 25 ): ): PUNCT iajs-3467 184 26 π‘₯ π‘₯ PROPN iajs-3467 184 27 ∈ ∈ PROPN iajs-3467 185 1 ℝ𝑛 ℝ𝑛 ADP iajs-3467 185 2 π‘Žπ‘›π‘‘ π‘Žπ‘›π‘‘ PROPN iajs-3467 185 3 𝑦 𝑦 PROPN iajs-3467 185 4 ∈ ∈ PROPN iajs-3467 185 5 { { PUNCT iajs-3467 185 6 βˆ’1 βˆ’1 NOUN iajs-3467 185 7 , , PUNCT iajs-3467 185 8 +1 +1 PROPN iajs-3467 185 9 } } PUNCT iajs-3467 185 10 } } PUNCT iajs-3467 185 11 repeat repeat VERB iajs-3467 185 12 until until ADP iajs-3467 185 13 convergence convergence NOUN iajs-3467 185 14 1 1 NUM iajs-3467 185 15 . . PUNCT iajs-3467 185 16 initial initial ADJ iajs-3467 185 17 value value NOUN iajs-3467 185 18 : : PUNCT iajs-3467 185 19 𝛽0 𝛽0 NOUN iajs-3467 185 20 . . PUNCT iajs-3467 186 1 2 2 X iajs-3467 186 2 . . X iajs-3467 186 3 for for ADP iajs-3467 186 4 𝑖 𝑖 PRON iajs-3467 186 5 = = SYM iajs-3467 186 6 1 1 NUM iajs-3467 186 7 , , PUNCT iajs-3467 186 8 … … PUNCT iajs-3467 186 9 , , PUNCT iajs-3467 186 10 k k NOUN iajs-3467 186 11 : : PUNCT iajs-3467 186 12 i. i. PROPN iajs-3467 186 13 compute compute PROPN iajs-3467 186 14 : : PUNCT iajs-3467 186 15 βˆ‡(𝛽𝑗 βˆ‡(𝛽𝑗 PROPN iajs-3467 186 16 ) ) PUNCT iajs-3467 186 17 = = SYM iajs-3467 186 18 πœ•π‘“(𝛽,𝛽0 πœ•π‘“(𝛽,𝛽0 NUM iajs-3467 186 19 ) ) PUNCT iajs-3467 186 20 πœ•π›½j πœ•π›½j PUNCT iajs-3467 187 1 = = PRON iajs-3467 188 1 𝛽j 𝛽j ADP iajs-3467 188 2 + + X iajs-3467 188 3 r r NOUN iajs-3467 188 4 βˆ‘ βˆ‘ SYM iajs-3467 188 5 πœ•πΏ(π‘₯𝑖,𝑦𝑖 πœ•πΏ(π‘₯𝑖,𝑦𝑖 PROPN iajs-3467 188 6 ) ) PUNCT iajs-3467 188 7 πœ•π›½j πœ•π›½j PUNCT iajs-3467 189 1 𝑛 𝑛 PRON iajs-3467 189 2 𝑖=1 𝑖=1 PROPN iajs-3467 189 3 , , PUNCT iajs-3467 189 4 where where SCONJ iajs-3467 189 5 𝑅 𝑅 PROPN iajs-3467 189 6 is be AUX iajs-3467 189 7 a a DET iajs-3467 189 8 regularization regularization NOUN iajs-3467 189 9 factor factor NOUN iajs-3467 189 10 . . PUNCT iajs-3467 190 1 ii ii PROPN iajs-3467 190 2 . . PUNCT iajs-3467 190 3 recompute recompute PROPN iajs-3467 190 4 𝛽 𝛽 PROPN iajs-3467 190 5 as as SCONJ iajs-3467 190 6 follows follow VERB iajs-3467 190 7 : : PUNCT iajs-3467 190 8 𝛽𝑗 𝛽𝑗 NUM iajs-3467 190 9 ← ← PROPN iajs-3467 190 10 𝛽𝑗 𝛽𝑗 INTJ iajs-3467 190 11 βˆ’ βˆ’ PROPN iajs-3467 190 12 πœ‚βˆ‡(𝛽𝑗 πœ‚βˆ‡(𝛽𝑗 PROPN iajs-3467 190 13 ) ) PUNCT iajs-3467 190 14 , , PUNCT iajs-3467 190 15 where where SCONJ iajs-3467 190 16 πœ‚ πœ‚ NOUN iajs-3467 190 17 is be AUX iajs-3467 190 18 the the DET iajs-3467 190 19 learning learning NOUN iajs-3467 190 20 rate rate NOUN iajs-3467 190 21 value value NOUN iajs-3467 190 22 . . PUNCT iajs-3467 191 1 iii iii X iajs-3467 191 2 . . PUNCT iajs-3467 191 3 βˆ‡jt(𝛽𝑑 βˆ‡jt(𝛽𝑑 PROPN iajs-3467 191 4 ) ) PUNCT iajs-3467 191 5 = = SYM iajs-3467 191 6 1 1 NUM iajs-3467 191 7 2 2 NUM iajs-3467 191 8 𝛽0 𝛽0 NOUN iajs-3467 191 9 𝑇𝛽0 𝑇𝛽0 NOUN iajs-3467 191 10 + + X iajs-3467 191 11 𝐾 𝐾 PROPN iajs-3467 191 12 Β· Β· PUNCT iajs-3467 191 13 𝑁 𝑁 PROPN iajs-3467 191 14 Β· Β· SYM iajs-3467 191 15 max max PROPN iajs-3467 191 16 ( ( PUNCT iajs-3467 191 17 0,1 0,1 NUM iajs-3467 191 18 βˆ’ βˆ’ NOUN iajs-3467 191 19 𝑦𝑖𝛽 𝑦𝑖𝛽 NOUN iajs-3467 191 20 𝑇π‘₯𝑖 𝑇π‘₯𝑖 PROPN iajs-3467 191 21 ) ) PUNCT iajs-3467 191 22 , , PUNCT iajs-3467 191 23 where where SCONJ iajs-3467 191 24 𝑁 𝑁 PROPN iajs-3467 191 25 is be AUX iajs-3467 191 26 the the DET iajs-3467 191 27 number number NOUN iajs-3467 191 28 of of ADP iajs-3467 191 29 training training NOUN iajs-3467 191 30 examples example NOUN iajs-3467 191 31 . . PUNCT iajs-3467 192 1 iv iv X iajs-3467 192 2 . . PUNCT iajs-3467 192 3 update update NOUN iajs-3467 192 4 𝛽 𝛽 PROPN iajs-3467 192 5 as as SCONJ iajs-3467 192 6 follows follow VERB iajs-3467 192 7 : : PUNCT iajs-3467 192 8 𝛽𝑑 𝛽𝑑 AUX iajs-3467 192 9 ← ← PROPN iajs-3467 192 10 π›½π‘‘βˆ’1 π›½π‘‘βˆ’1 VERB iajs-3467 192 11 βˆ’ βˆ’ PROPN iajs-3467 192 12 π›Ύπ‘‘βˆ‡jt(π›½π‘‘βˆ’1 π›Ύπ‘‘βˆ‡jt(π›½π‘‘βˆ’1 PROPN iajs-3467 192 13 ) ) PUNCT iajs-3467 192 14 . . PUNCT iajs-3467 193 1 3 3 X iajs-3467 193 2 . . X iajs-3467 193 3 repeat repeat NOUN iajs-3467 193 4 ( ( PUNCT iajs-3467 193 5 π‘₯𝑖 π‘₯𝑖 PROPN iajs-3467 193 6 , , PUNCT iajs-3467 193 7 𝑦𝑖 𝑦𝑖 PROPN iajs-3467 193 8 ) ) PUNCT iajs-3467 193 9 to to PART iajs-3467 193 10 make make VERB iajs-3467 193 11 a a DET iajs-3467 193 12 full full ADJ iajs-3467 193 13 dataset dataset NOUN iajs-3467 193 14 and and CCONJ iajs-3467 193 15 take take VERB iajs-3467 193 16 the the DET iajs-3467 193 17 derivative derivative NOUN iajs-3467 193 18 of of ADP iajs-3467 193 19 the the DET iajs-3467 193 20 svm svm ADJ iajs-3467 193 21 objective objective NOUN iajs-3467 193 22 at at ADP iajs-3467 193 23 the the DET iajs-3467 193 24 current current NOUN iajs-3467 193 25 π›½π‘‘βˆ’1 π›½π‘‘βˆ’1 NOUN iajs-3467 193 26 to to PART iajs-3467 193 27 be be AUX iajs-3467 193 28 βˆ‡jt(π›½π‘‘βˆ’1 βˆ‡jt(π›½π‘‘βˆ’1 VERB iajs-3467 193 29 ) ) PUNCT iajs-3467 193 30 . . PUNCT iajs-3467 194 1 4 4 X iajs-3467 194 2 . . X iajs-3467 194 3 return return VERB iajs-3467 194 4 final final ADJ iajs-3467 194 5 𝛽. 𝛽. NOUN iajs-3467 194 6 ihjpas ihjpas PROPN iajs-3467 194 7 . . PUNCT iajs-3467 195 1 37 37 NUM iajs-3467 195 2 ( ( PUNCT iajs-3467 195 3 1 1 NUM iajs-3467 195 4 ) ) PUNCT iajs-3467 195 5 2024 2024 NUM iajs-3467 195 6 421 421 NUM iajs-3467 195 7 6 6 NUM iajs-3467 195 8 . . PUNCT iajs-3467 196 1 simulation simulation NOUN iajs-3467 196 2 studies study NOUN iajs-3467 196 3 to to PART iajs-3467 196 4 test test VERB iajs-3467 196 5 the the DET iajs-3467 196 6 gd gd NOUN iajs-3467 196 7 - - PUNCT iajs-3467 196 8 svm svm ADJ iajs-3467 196 9 and and CCONJ iajs-3467 196 10 esgd esgd NOUN iajs-3467 196 11 - - PUNCT iajs-3467 196 12 svm svm ADJ iajs-3467 196 13 methods method NOUN iajs-3467 196 14 , , PUNCT iajs-3467 196 15 two two NUM iajs-3467 196 16 simulation simulation NOUN iajs-3467 196 17 datasets dataset NOUN iajs-3467 196 18 with with ADP iajs-3467 196 19 100 100 NUM iajs-3467 196 20 and and CCONJ iajs-3467 196 21 200 200 NUM iajs-3467 196 22 observations observation NOUN iajs-3467 196 23 are be AUX iajs-3467 196 24 created create VERB iajs-3467 196 25 . . PUNCT iajs-3467 197 1 each each DET iajs-3467 197 2 dataset dataset NOUN iajs-3467 197 3 contains contain VERB iajs-3467 197 4 a a DET iajs-3467 197 5 set set NOUN iajs-3467 197 6 of of ADP iajs-3467 197 7 variables variable NOUN iajs-3467 197 8 and and CCONJ iajs-3467 197 9 a a DET iajs-3467 197 10 response response NOUN iajs-3467 197 11 comprising comprise VERB iajs-3467 197 12 two two NUM iajs-3467 197 13 classes class NOUN iajs-3467 197 14 . . PUNCT iajs-3467 198 1 in in ADP iajs-3467 198 2 real real ADJ iajs-3467 198 3 life life NOUN iajs-3467 198 4 , , PUNCT iajs-3467 198 5 it it PRON iajs-3467 198 6 is be AUX iajs-3467 198 7 not not PART iajs-3467 198 8 known know VERB iajs-3467 198 9 whether whether SCONJ iajs-3467 198 10 the the DET iajs-3467 198 11 data data NOUN iajs-3467 198 12 sets set NOUN iajs-3467 198 13 are be AUX iajs-3467 198 14 linearly linearly ADV iajs-3467 198 15 or or CCONJ iajs-3467 198 16 non non ADJ iajs-3467 198 17 - - ADJ iajs-3467 198 18 linearly linearly ADV iajs-3467 198 19 separable separable ADJ iajs-3467 198 20 , , PUNCT iajs-3467 198 21 so so ADV iajs-3467 198 22 complex complex ADJ iajs-3467 198 23 data data NOUN iajs-3467 198 24 sets set NOUN iajs-3467 198 25 ( ( PUNCT iajs-3467 198 26 non non ADJ iajs-3467 198 27 - - ADJ iajs-3467 198 28 linearly linearly ADV iajs-3467 198 29 separable separable NOUN iajs-3467 198 30 ) ) PUNCT iajs-3467 198 31 are be AUX iajs-3467 198 32 generated generate VERB iajs-3467 198 33 . . PUNCT iajs-3467 199 1 gd gd ADJ iajs-3467 199 2 - - PUNCT iajs-3467 199 3 svm svm ADJ iajs-3467 199 4 and and CCONJ iajs-3467 199 5 esgd esgd NOUN iajs-3467 199 6 - - PUNCT iajs-3467 199 7 svm svm PROPN iajs-3467 199 8 are be AUX iajs-3467 199 9 applied apply VERB iajs-3467 199 10 to to ADP iajs-3467 199 11 the the DET iajs-3467 199 12 two two NUM iajs-3467 199 13 data data NOUN iajs-3467 199 14 sets set NOUN iajs-3467 199 15 by by ADP iajs-3467 199 16 using use VERB iajs-3467 199 17 different different ADJ iajs-3467 199 18 types type NOUN iajs-3467 199 19 of of ADP iajs-3467 199 20 kernels kernel NOUN iajs-3467 199 21 . . PUNCT iajs-3467 200 1 in in ADP iajs-3467 200 2 both both DET iajs-3467 200 3 methods method NOUN iajs-3467 200 4 , , PUNCT iajs-3467 200 5 the the DET iajs-3467 200 6 mlp mlp PROPN iajs-3467 200 7 kernel kernel PROPN iajs-3467 200 8 provides provide VERB iajs-3467 200 9 the the DET iajs-3467 200 10 best good ADJ iajs-3467 200 11 accuracy accuracy NOUN iajs-3467 200 12 in in ADP iajs-3467 200 13 the the DET iajs-3467 200 14 two two NUM iajs-3467 200 15 datasets dataset NOUN iajs-3467 200 16 [ [ X iajs-3467 200 17 23,24 23,24 NOUN iajs-3467 200 18 ] ] PUNCT iajs-3467 200 19 . . PUNCT iajs-3467 201 1 in in ADP iajs-3467 201 2 table table NOUN iajs-3467 201 3 1 1 NUM iajs-3467 201 4 , , PUNCT iajs-3467 201 5 gd gd NOUN iajs-3467 201 6 - - PUNCT iajs-3467 201 7 svm svm ADJ iajs-3467 201 8 and and CCONJ iajs-3467 201 9 esgd esgd NOUN iajs-3467 201 10 - - PUNCT iajs-3467 201 11 svm svm PROPN iajs-3467 201 12 are be AUX iajs-3467 201 13 compared compare VERB iajs-3467 201 14 in in ADP iajs-3467 201 15 terms term NOUN iajs-3467 201 16 of of ADP iajs-3467 201 17 the the DET iajs-3467 201 18 best good ADJ iajs-3467 201 19 value value NOUN iajs-3467 201 20 of of ADP iajs-3467 201 21 𝑝 𝑝 NOUN iajs-3467 201 22 , , PUNCT iajs-3467 201 23 the the DET iajs-3467 201 24 number number NOUN iajs-3467 201 25 of of ADP iajs-3467 201 26 support support NOUN iajs-3467 201 27 vectors vector NOUN iajs-3467 201 28 , , PUNCT iajs-3467 201 29 using use VERB iajs-3467 201 30 sensitivity sensitivity NOUN iajs-3467 201 31 ( ( PUNCT iajs-3467 201 32 also also ADV iajs-3467 201 33 called call VERB iajs-3467 201 34 truepositive truepositive ADJ iajs-3467 201 35 rate rate NOUN iajs-3467 201 36 ) ) PUNCT iajs-3467 201 37 and and CCONJ iajs-3467 201 38 specificity specificity NOUN iajs-3467 201 39 ( ( PUNCT iajs-3467 201 40 also also ADV iajs-3467 201 41 called call VERB iajs-3467 201 42 true true ADJ iajs-3467 201 43 - - PUNCT iajs-3467 201 44 negative negative ADJ iajs-3467 201 45 rate rate NOUN iajs-3467 201 46 ) ) PUNCT iajs-3467 202 1 [ [ X iajs-3467 202 2 25,26 25,26 X iajs-3467 202 3 ] ] PUNCT iajs-3467 202 4 . . PUNCT iajs-3467 203 1 table table NOUN iajs-3467 203 2 1 1 NUM iajs-3467 203 3 . . PUNCT iajs-3467 203 4 comparison comparison NOUN iajs-3467 203 5 between between ADP iajs-3467 203 6 gd gd NOUN iajs-3467 203 7 - - PUNCT iajs-3467 203 8 svm svm PROPN iajs-3467 203 9 & & CCONJ iajs-3467 203 10 esgd esgd PROPN iajs-3467 203 11 - - PUNCT iajs-3467 203 12 svm svm PROPN iajs-3467 203 13 for for ADP iajs-3467 203 14 a a DET iajs-3467 203 15 data datum NOUN iajs-3467 203 16 set set VERB iajs-3467 203 17 with with ADP iajs-3467 203 18 100 100 NUM iajs-3467 203 19 and and CCONJ iajs-3467 203 20 200 200 NUM iajs-3467 203 21 observations observation NOUN iajs-3467 203 22 . . PUNCT iajs-3467 204 1 sample sample NOUN iajs-3467 204 2 size size NOUN iajs-3467 204 3 method method NOUN iajs-3467 204 4 kernel kernel PROPN iajs-3467 204 5 type type NOUN iajs-3467 204 6 best good ADJ iajs-3467 204 7 𝑲 𝑲 DET iajs-3467 204 8 value value NOUN iajs-3467 204 9 number number NOUN iajs-3467 204 10 of of ADP iajs-3467 204 11 support support NOUN iajs-3467 204 12 vectors vector NOUN iajs-3467 204 13 sensitivity% sensitivity% PUNCT iajs-3467 204 14 specificity% specificity% PUNCT iajs-3467 204 15 accuracy accuracy NOUN iajs-3467 204 16 100 100 NUM iajs-3467 204 17 observations observation NOUN iajs-3467 204 18 gd gd NOUN iajs-3467 204 19 - - PUNCT iajs-3467 204 20 svm svm NOUN iajs-3467 204 21 linear linear NOUN iajs-3467 204 22 0.78 0.78 NUM iajs-3467 204 23 13 13 NUM iajs-3467 204 24 82.30 82.30 NUM iajs-3467 204 25 89.43 89.43 NUM iajs-3467 204 26 86.43 86.43 NUM iajs-3467 204 27 polynomial polynomial NOUN iajs-3467 204 28 0.45 0.45 NUM iajs-3467 204 29 31 31 NUM iajs-3467 204 30 94.40 94.40 NUM iajs-3467 204 31 92.58 92.58 NUM iajs-3467 204 32 93.56 93.56 NUM iajs-3467 204 33 rbf rbf PROPN iajs-3467 204 34 0.32 0.32 NUM iajs-3467 204 35 19 19 NUM iajs-3467 204 36 92.74 92.74 NUM iajs-3467 204 37 94.01 94.01 NUM iajs-3467 204 38 93.01 93.01 NUM iajs-3467 204 39 mlp mlp NOUN iajs-3467 204 40 0.23 0.23 NUM iajs-3467 204 41 22 22 NUM iajs-3467 204 42 90.95 90.95 NUM iajs-3467 204 43 94.33 94.33 NUM iajs-3467 204 44 92.33 92.33 NUM iajs-3467 204 45 esgdsvm esgdsvm NOUN iajs-3467 204 46 linear linear PROPN iajs-3467 204 47 0.96 0.96 NUM iajs-3467 204 48 12 12 NUM iajs-3467 204 49 92.30 92.30 NUM iajs-3467 204 50 97.43 97.43 NUM iajs-3467 204 51 93.45 93.45 NUM iajs-3467 204 52 polynomial polynomial ADJ iajs-3467 204 53 0.55 0.55 NUM iajs-3467 204 54 19 19 NUM iajs-3467 204 55 95.40 95.40 NUM iajs-3467 204 56 96.53 96.53 NUM iajs-3467 204 57 95.53 95.53 NUM iajs-3467 204 58 rbf rbf PROPN iajs-3467 204 59 0.37 0.37 NUM iajs-3467 204 60 15 15 NUM iajs-3467 204 61 96.74 96.74 NUM iajs-3467 204 62 97.01 97.01 NUM iajs-3467 204 63 96.53 96.53 NUM iajs-3467 204 64 mlp mlp NOUN iajs-3467 204 65 0.33 0.33 NUM iajs-3467 204 66 16 16 NUM iajs-3467 204 67 98.95 98.95 NUM iajs-3467 204 68 96.33 96.33 NUM iajs-3467 204 69 97.33 97.33 NUM iajs-3467 204 70 200 200 NUM iajs-3467 204 71 observations observation NOUN iajs-3467 204 72 gdsvm gdsvm VERB iajs-3467 204 73 linear linear PROPN iajs-3467 204 74 0.83 0.83 NUM iajs-3467 204 75 23 23 NUM iajs-3467 204 76 81.31 81.31 NUM iajs-3467 204 77 87.43 87.43 NUM iajs-3467 204 78 85.43 85.43 NUM iajs-3467 204 79 polynomial polynomial ADJ iajs-3467 204 80 0.26 0.26 NUM iajs-3467 204 81 41 41 NUM iajs-3467 204 82 89.44 89.44 NUM iajs-3467 204 83 91.55 91.55 NUM iajs-3467 204 84 90.54 90.54 NUM iajs-3467 204 85 rbf rbf PROPN iajs-3467 204 86 0.39 0.39 NUM iajs-3467 204 87 28 28 NUM iajs-3467 204 88 89.74 89.74 NUM iajs-3467 204 89 84.01 84.01 NUM iajs-3467 204 90 87.01 87.01 NUM iajs-3467 204 91 mlp mlp NOUN iajs-3467 204 92 0.24 0.24 NUM iajs-3467 204 93 33 33 NUM iajs-3467 204 94 89.94 89.94 NUM iajs-3467 204 95 91.73 91.73 NUM iajs-3467 204 96 90.72 90.72 NUM iajs-3467 204 97 esgdsvm esgdsvm NOUN iajs-3467 204 98 linear linear PROPN iajs-3467 204 99 0.92 0.92 NUM iajs-3467 204 100 14 14 NUM iajs-3467 204 101 80.33 80.33 NUM iajs-3467 204 102 83.83 83.83 NUM iajs-3467 204 103 82.81 82.81 NUM iajs-3467 204 104 polynomial polynomial ADJ iajs-3467 204 105 0.57 0.57 NUM iajs-3467 204 106 29 29 NUM iajs-3467 204 107 87.43 87.43 NUM iajs-3467 204 108 88.58 88.58 NUM iajs-3467 204 109 86.54 86.54 NUM iajs-3467 204 110 rbf rbf PROPN iajs-3467 204 111 0.48 0.48 NUM iajs-3467 204 112 25 25 NUM iajs-3467 204 113 88.44 88.44 NUM iajs-3467 204 114 87.97 87.97 NUM iajs-3467 204 115 87.99 87.99 NUM iajs-3467 204 116 mlp mlp NOUN iajs-3467 204 117 0.37 0.37 NUM iajs-3467 204 118 28 28 NUM iajs-3467 204 119 86.41 86.41 NUM iajs-3467 204 120 85.96 85.96 NUM iajs-3467 204 121 85.66 85.66 NUM iajs-3467 204 122 7 7 NUM iajs-3467 204 123 . . PUNCT iajs-3467 205 1 real real ADV iajs-3467 205 2 dataset dataset VERB iajs-3467 205 3 the the DET iajs-3467 205 4 modification modification NOUN iajs-3467 205 5 of of ADP iajs-3467 205 6 the the DET iajs-3467 205 7 svm svm NOUN iajs-3467 205 8 using use VERB iajs-3467 205 9 stochastic stochastic ADJ iajs-3467 205 10 gradient gradient ADJ iajs-3467 205 11 descent descent NOUN iajs-3467 205 12 is be AUX iajs-3467 205 13 the the DET iajs-3467 205 14 main main ADJ iajs-3467 205 15 topic topic NOUN iajs-3467 205 16 of of ADP iajs-3467 205 17 this this DET iajs-3467 205 18 paper paper NOUN iajs-3467 205 19 . . PUNCT iajs-3467 206 1 for for ADP iajs-3467 206 2 real real ADJ iajs-3467 206 3 data datum NOUN iajs-3467 206 4 applications application NOUN iajs-3467 206 5 , , PUNCT iajs-3467 206 6 we we PRON iajs-3467 206 7 have have AUX iajs-3467 206 8 used use VERB iajs-3467 206 9 south south ADJ iajs-3467 206 10 african african ADJ iajs-3467 206 11 heart heart NOUN iajs-3467 206 12 disease disease NOUN iajs-3467 206 13 data datum NOUN iajs-3467 206 14 . . PUNCT iajs-3467 207 1 the the DET iajs-3467 207 2 dataset dataset NOUN iajs-3467 207 3 was be AUX iajs-3467 207 4 downloaded download VERB iajs-3467 207 5 from from ADP iajs-3467 207 6 [ [ X iajs-3467 207 7 18 18 NUM iajs-3467 207 8 ] ] PUNCT iajs-3467 207 9 . . PUNCT iajs-3467 208 1 in in ADP iajs-3467 208 2 this this DET iajs-3467 208 3 dataset dataset NOUN iajs-3467 208 4 , , PUNCT iajs-3467 208 5 a a DET iajs-3467 208 6 historical historical ADJ iajs-3467 208 7 sample sample NOUN iajs-3467 208 8 of of ADP iajs-3467 208 9 men man NOUN iajs-3467 208 10 in in ADP iajs-3467 208 11 the the DET iajs-3467 208 12 western western ADJ iajs-3467 208 13 cape cape NOUN iajs-3467 208 14 of of ADP iajs-3467 208 15 south south PROPN iajs-3467 208 16 africa africa PROPN iajs-3467 208 17 , , PUNCT iajs-3467 208 18 a a DET iajs-3467 208 19 region region NOUN iajs-3467 208 20 with with ADP iajs-3467 208 21 a a DET iajs-3467 208 22 high high ADJ iajs-3467 208 23 incidence incidence NOUN iajs-3467 208 24 of of ADP iajs-3467 208 25 cardiovascular cardiovascular ADJ iajs-3467 208 26 disease disease NOUN iajs-3467 208 27 , , PUNCT iajs-3467 208 28 is be AUX iajs-3467 208 29 described describe VERB iajs-3467 208 30 . . PUNCT iajs-3467 209 1 the the DET iajs-3467 209 2 following follow VERB iajs-3467 209 3 patient patient ADJ iajs-3467 209 4 characteristics characteristic NOUN iajs-3467 209 5 were be AUX iajs-3467 209 6 recorded record VERB iajs-3467 209 7 for for ADP iajs-3467 209 8 each each DET iajs-3467 209 9 high high ADJ iajs-3467 209 10 - - PUNCT iajs-3467 209 11 risk risk NOUN iajs-3467 209 12 individual individual NOUN iajs-3467 209 13 : : PUNCT iajs-3467 209 14 factors factor NOUN iajs-3467 209 15 such such ADJ iajs-3467 209 16 as as ADP iajs-3467 209 17 age age NOUN iajs-3467 209 18 , , PUNCT iajs-3467 209 19 type type VERB iajs-3467 209 20 a a DET iajs-3467 209 21 behavior behavior NOUN iajs-3467 209 22 , , PUNCT iajs-3467 209 23 family family NOUN iajs-3467 209 24 history history NOUN iajs-3467 209 25 of of ADP iajs-3467 209 26 heart heart NOUN iajs-3467 209 27 disease disease NOUN iajs-3467 209 28 , , PUNCT iajs-3467 209 29 systolic systolic ADJ iajs-3467 209 30 blood blood NOUN iajs-3467 209 31 pressure pressure NOUN iajs-3467 209 32 , , PUNCT iajs-3467 209 33 cumulative cumulative ADJ iajs-3467 209 34 cigarette cigarette NOUN iajs-3467 209 35 consumption consumption NOUN iajs-3467 209 36 , , PUNCT iajs-3467 209 37 lowdensity lowdensity NOUN iajs-3467 209 38 lipoprotein lipoprotein ADJ iajs-3467 209 39 cholesterol cholesterol NOUN iajs-3467 209 40 , , PUNCT iajs-3467 209 41 body body NOUN iajs-3467 209 42 fat fat ADJ iajs-3467 209 43 percentage percentage NOUN iajs-3467 209 44 and and CCONJ iajs-3467 209 45 obesity obesity NOUN iajs-3467 209 46 . . PUNCT iajs-3467 210 1 the the DET iajs-3467 210 2 total total ADJ iajs-3467 210 3 number number NOUN iajs-3467 210 4 of of ADP iajs-3467 210 5 samples sample NOUN iajs-3467 210 6 in in ADP iajs-3467 210 7 this this DET iajs-3467 210 8 data data NOUN iajs-3467 210 9 collection collection NOUN iajs-3467 210 10 is be AUX iajs-3467 210 11 462 462 NUM iajs-3467 210 12 . . PUNCT iajs-3467 211 1 obesity obesity NOUN iajs-3467 211 2 refers refer VERB iajs-3467 211 3 to to ADP iajs-3467 211 4 a a DET iajs-3467 211 5 high high ADJ iajs-3467 211 6 percentage percentage NOUN iajs-3467 211 7 of of ADP iajs-3467 211 8 body body NOUN iajs-3467 211 9 fat fat ADJ iajs-3467 211 10 , , PUNCT iajs-3467 211 11 while while SCONJ iajs-3467 211 12 obesity obesity NOUN iajs-3467 211 13 is be AUX iajs-3467 211 14 defined define VERB iajs-3467 211 15 by by ADP iajs-3467 211 16 a a DET iajs-3467 211 17 high high ADJ iajs-3467 211 18 weight weight NOUN iajs-3467 211 19 - - PUNCT iajs-3467 211 20 to to ADP iajs-3467 211 21 - - PUNCT iajs-3467 211 22 height height NOUN iajs-3467 211 23 ratio ratio NOUN iajs-3467 211 24 ( ( PUNCT iajs-3467 211 25 body body NOUN iajs-3467 211 26 mass mass PROPN iajs-3467 211 27 index index PROPN iajs-3467 211 28 , , PUNCT iajs-3467 211 29 bmi bmi PROPN iajs-3467 211 30 ) ) PUNCT iajs-3467 212 1 [ [ X iajs-3467 212 2 27 27 NUM iajs-3467 212 3 ] ] SYM iajs-3467 212 4 excessive excessive ADJ iajs-3467 212 5 antagonism antagonism NOUN iajs-3467 212 6 , , PUNCT iajs-3467 212 7 aggression aggression NOUN iajs-3467 212 8 , , PUNCT iajs-3467 212 9 and and CCONJ iajs-3467 212 10 competitiveness competitiveness NOUN iajs-3467 212 11 are be AUX iajs-3467 212 12 hallmarks hallmark NOUN iajs-3467 212 13 of of ADP iajs-3467 212 14 the the DET iajs-3467 212 15 type type NOUN iajs-3467 212 16 a a DET iajs-3467 212 17 personality personality NOUN iajs-3467 212 18 . . PUNCT iajs-3467 213 1 we we PRON iajs-3467 213 2 will will AUX iajs-3467 213 3 see see VERB iajs-3467 213 4 if if SCONJ iajs-3467 213 5 we we PRON iajs-3467 213 6 can can AUX iajs-3467 213 7 extrapolate extrapolate VERB iajs-3467 213 8 ldl ldl NOUN iajs-3467 213 9 from from ADP iajs-3467 213 10 the the DET iajs-3467 213 11 available available ADJ iajs-3467 213 12 data datum NOUN iajs-3467 213 13 . . PUNCT iajs-3467 214 1 since since SCONJ iajs-3467 214 2 low low ADJ iajs-3467 214 3 - - PUNCT iajs-3467 214 4 density density NOUN iajs-3467 214 5 lipoprotein lipoprotein ADJ iajs-3467 214 6 ( ( PUNCT iajs-3467 214 7 ldl ldl NOUN iajs-3467 214 8 ) ) PUNCT iajs-3467 214 9 cholesterol cholesterol NOUN iajs-3467 214 10 is be AUX iajs-3467 214 11 the the DET iajs-3467 214 12 " " PUNCT iajs-3467 214 13 bad bad ADJ iajs-3467 214 14 " " PUNCT iajs-3467 214 15 cholesterol cholesterol NOUN iajs-3467 214 16 , , PUNCT iajs-3467 214 17 elevated elevated ADJ iajs-3467 214 18 levels level NOUN iajs-3467 214 19 are be AUX iajs-3467 214 20 thought think VERB iajs-3467 214 21 to to PART iajs-3467 214 22 be be AUX iajs-3467 214 23 associated associate VERB iajs-3467 214 24 with with ADP iajs-3467 214 25 obesity obesity NOUN iajs-3467 214 26 and and CCONJ iajs-3467 214 27 adiposity adiposity NOUN iajs-3467 214 28 . . PUNCT iajs-3467 215 1 the the DET iajs-3467 215 2 aim aim NOUN iajs-3467 215 3 is be AUX iajs-3467 215 4 to to PART iajs-3467 215 5 create create VERB iajs-3467 215 6 a a DET iajs-3467 215 7 predictive predictive ADJ iajs-3467 215 8 model model NOUN iajs-3467 215 9 for for ADP iajs-3467 215 10 ldl ldl NOUN iajs-3467 215 11 by by ADP iajs-3467 215 12 selecting select VERB iajs-3467 215 13 the the DET iajs-3467 215 14 most most ADV iajs-3467 215 15 important important ADJ iajs-3467 215 16 factors factor NOUN iajs-3467 215 17 [ [ X iajs-3467 215 18 28 28 NUM iajs-3467 215 19 ] ] PUNCT iajs-3467 215 20 . . PUNCT iajs-3467 216 1 in in ADP iajs-3467 216 2 table table NOUN iajs-3467 216 3 ( ( PUNCT iajs-3467 216 4 2 2 NUM iajs-3467 216 5 ) ) PUNCT iajs-3467 216 6 , , PUNCT iajs-3467 216 7 we we PRON iajs-3467 216 8 examine examine VERB iajs-3467 216 9 the the DET iajs-3467 216 10 pearson pearson PROPN iajs-3467 216 11 correlation correlation NOUN iajs-3467 216 12 coefficient coefficient NOUN iajs-3467 216 13 between between ADP iajs-3467 216 14 the the DET iajs-3467 216 15 groups group NOUN iajs-3467 216 16 . . PUNCT iajs-3467 217 1 it it PRON iajs-3467 217 2 can can AUX iajs-3467 217 3 be be AUX iajs-3467 217 4 seen see VERB iajs-3467 217 5 that that SCONJ iajs-3467 217 6 there there PRON iajs-3467 217 7 is be VERB iajs-3467 217 8 a a DET iajs-3467 217 9 high high ADJ iajs-3467 217 10 significant significant ADJ iajs-3467 217 11 correlation correlation NOUN iajs-3467 217 12 ( ( PUNCT iajs-3467 217 13 marked mark VERB iajs-3467 217 14 with with ADP iajs-3467 217 15 * * PROPN iajs-3467 217 16 * * PUNCT iajs-3467 217 17 ) ) PUNCT iajs-3467 217 18 , , PUNCT iajs-3467 217 19 a a DET iajs-3467 217 20 significant significant ADJ iajs-3467 217 21 correlation correlation NOUN iajs-3467 217 22 ( ( PUNCT iajs-3467 217 23 marked mark VERB iajs-3467 217 24 with with ADP iajs-3467 217 25 * * PUNCT iajs-3467 217 26 ) ) PUNCT iajs-3467 217 27 and and CCONJ iajs-3467 217 28 a a DET iajs-3467 217 29 weak weak ADJ iajs-3467 217 30 correlation correlation NOUN iajs-3467 217 31 between between ADP iajs-3467 217 32 the the DET iajs-3467 217 33 variables variable NOUN iajs-3467 217 34 [ [ X iajs-3467 217 35 29,30 29,30 NUM iajs-3467 217 36 ] ] PUNCT iajs-3467 217 37 . . PUNCT iajs-3467 218 1 ihjpas ihjpas PROPN iajs-3467 218 2 . . PUNCT iajs-3467 219 1 37 37 NUM iajs-3467 219 2 ( ( PUNCT iajs-3467 219 3 1 1 NUM iajs-3467 219 4 ) ) PUNCT iajs-3467 219 5 2024 2024 NUM iajs-3467 219 6 422 422 NUM iajs-3467 219 7 table table NOUN iajs-3467 219 8 2 2 NUM iajs-3467 219 9 . . PUNCT iajs-3467 219 10 pearson pearson PROPN iajs-3467 219 11 correlation correlation NOUN iajs-3467 219 12 matrix matrix NOUN iajs-3467 219 13 between between ADP iajs-3467 219 14 variables variable NOUN iajs-3467 219 15 of of ADP iajs-3467 219 16 heart heart NOUN iajs-3467 219 17 disease disease NOUN iajs-3467 219 18 dataset dataset VERB iajs-3467 219 19 pearson pearson PROPN iajs-3467 219 20 correlation correlation PROPN iajs-3467 219 21 sbp sbp NOUN iajs-3467 219 22 tobacco tobacco NOUN iajs-3467 219 23 ldl ldl NOUN iajs-3467 219 24 adiposity adiposity NOUN iajs-3467 219 25 typea typea NOUN iajs-3467 219 26 obesity obesity NOUN iajs-3467 219 27 alcohol alcohol NOUN iajs-3467 219 28 tobacco tobacco NOUN iajs-3467 219 29 r r NOUN iajs-3467 219 30 p p NOUN iajs-3467 219 31 0.212 0.212 NUM iajs-3467 219 32 0.092 0.092 NUM iajs-3467 219 33 ldl ldl NOUN iajs-3467 219 34 r r NOUN iajs-3467 219 35 p p NOUN iajs-3467 219 36 0.158 0.158 NUM iajs-3467 219 37 0.958 0.958 NUM iajs-3467 219 38 0.159 0.159 NUM iajs-3467 219 39 0.203 0.203 NUM iajs-3467 219 40 adiposity adiposity NOUN iajs-3467 219 41 r r NOUN iajs-3467 219 42 p p NOUN iajs-3467 219 43 0.357 0.357 NUM iajs-3467 219 44 * * PUNCT iajs-3467 219 45 0.043 0.043 NUM iajs-3467 219 46 0.287 0.287 NUM iajs-3467 219 47 0.349 0.349 NUM iajs-3467 219 48 0.440 0.440 NUM iajs-3467 219 49 * * SYM iajs-3467 219 50 * * SYM iajs-3467 219 51 0.002 0.002 NUM iajs-3467 219 52 typea typea NOUN iajs-3467 219 53 r r NOUN iajs-3467 219 54 p p NOUN iajs-3467 219 55 -0.057 -0.057 NOUN iajs-3467 219 56 0.945 0.945 NUM iajs-3467 219 57 -0.015 -0.015 NUM iajs-3467 219 58 0.503 0.503 NUM iajs-3467 219 59 0.440 0.440 NUM iajs-3467 219 60 * * NOUN iajs-3467 219 61 0.039 0.039 NUM iajs-3467 219 62 -0.043 -0.043 NUM iajs-3467 219 63 0.203 0.203 NUM iajs-3467 219 64 obesity obesity NOUN iajs-3467 219 65 r r NOUN iajs-3467 219 66 p p NOUN iajs-3467 219 67 0.238 0.238 NUM iajs-3467 219 68 * * PROPN iajs-3467 219 69 0.043 0.043 NUM iajs-3467 219 70 0.125 0.125 NUM iajs-3467 219 71 0.349 0.349 NUM iajs-3467 219 72 0.331 0.331 NUM iajs-3467 219 73 * * NOUN iajs-3467 220 1 0.049 0.049 NUM iajs-3467 220 2 0.717 0.717 NUM iajs-3467 220 3 * * PUNCT iajs-3467 220 4 * * PUNCT iajs-3467 220 5 0.014 0.014 NUM iajs-3467 220 6 0.074 0.074 NUM iajs-3467 220 7 0.249 0.249 NUM iajs-3467 220 8 alcohol alcohol NOUN iajs-3467 220 9 r r NOUN iajs-3467 220 10 p p NOUN iajs-3467 220 11 0.140 0.140 NUM iajs-3467 220 12 0.4l3 0.4l3 NUM iajs-3467 220 13 0.201 0.201 NUM iajs-3467 220 14 0.459 0.459 NUM iajs-3467 220 15 -0.033 -0.033 PUNCT iajs-3467 220 16 0.953 0.953 NUM iajs-3467 220 17 0.123 0.123 NUM iajs-3467 220 18 0.034 0.034 NUM iajs-3467 220 19 0.039 0.039 NUM iajs-3467 220 20 0.254 0.254 NUM iajs-3467 220 21 0.052 0.052 NUM iajs-3467 220 22 0.359 0.359 NUM iajs-3467 220 23 age age NOUN iajs-3467 220 24 r r NOUN iajs-3467 220 25 p p NOUN iajs-3467 220 26 0.389 0.389 NUM iajs-3467 220 27 * * NOUN iajs-3467 220 28 0.042 0.042 NUM iajs-3467 220 29 0.450 0.450 NUM iajs-3467 220 30 * * PUNCT iajs-3467 220 31 * * PUNCT iajs-3467 220 32 0.011 0.011 NUM iajs-3467 220 33 0.312 0.312 NUM iajs-3467 220 34 0.059 0.059 NUM iajs-3467 220 35 0.626 0.626 NUM iajs-3467 220 36 * * PUNCT iajs-3467 220 37 * * PROPN iajs-3467 220 38 0.023 0.023 NUM iajs-3467 220 39 -0.103 -0.103 NUM iajs-3467 220 40 0.539 0.539 NUM iajs-3467 220 41 0.692 0.692 NUM iajs-3467 220 42 * * PUNCT iajs-3467 220 43 * * PUNCT iajs-3467 220 44 0.024 0.024 NUM iajs-3467 220 45 0.101 0.101 NUM iajs-3467 220 46 0.239 0.239 NUM iajs-3467 220 47 r r NOUN iajs-3467 220 48 : : PUNCT iajs-3467 220 49 pearson pearson NOUN iajs-3467 220 50 correlation correlation NOUN iajs-3467 220 51 , , PUNCT iajs-3467 220 52 p p X iajs-3467 220 53 : : PUNCT iajs-3467 220 54 p p ADJ iajs-3467 220 55 value value NOUN iajs-3467 220 56 * * PUNCT iajs-3467 220 57 * * ADJ iajs-3467 220 58 high high ADJ iajs-3467 220 59 significant significant ADJ iajs-3467 220 60 correlation correlation NOUN iajs-3467 220 61 between between ADP iajs-3467 220 62 variables variable NOUN iajs-3467 220 63 ( ( PUNCT iajs-3467 220 64 p p NOUN iajs-3467 220 65 value value NOUN iajs-3467 220 66 < < X iajs-3467 220 67 0.01 0.01 NUM iajs-3467 220 68 ) ) PUNCT iajs-3467 220 69 . . PUNCT iajs-3467 221 1 * * PUNCT iajs-3467 221 2 significant significant ADJ iajs-3467 221 3 correlation correlation NOUN iajs-3467 221 4 between between ADP iajs-3467 221 5 variables variable NOUN iajs-3467 221 6 ( ( PUNCT iajs-3467 221 7 p p NOUN iajs-3467 221 8 value value NOUN iajs-3467 221 9 < < X iajs-3467 221 10 0.05 0.05 NUM iajs-3467 221 11 ) ) PUNCT iajs-3467 221 12 . . PUNCT iajs-3467 222 1 the the DET iajs-3467 222 2 estimated estimate VERB iajs-3467 222 3 coefficient coefficient NOUN iajs-3467 222 4 for for ADP iajs-3467 222 5 the the DET iajs-3467 222 6 model model NOUN iajs-3467 222 7 was be AUX iajs-3467 222 8 calculated calculate VERB iajs-3467 222 9 in in ADP iajs-3467 222 10 table table NOUN iajs-3467 222 11 ( ( PUNCT iajs-3467 222 12 3 3 NUM iajs-3467 222 13 ) ) PUNCT iajs-3467 222 14 , , PUNCT iajs-3467 222 15 some some DET iajs-3467 222 16 variables variable NOUN iajs-3467 222 17 were be AUX iajs-3467 222 18 highly highly ADV iajs-3467 222 19 significant significant ADJ iajs-3467 222 20 since since SCONJ iajs-3467 222 21 p p NOUN iajs-3467 222 22 value value NOUN iajs-3467 222 23 was be AUX iajs-3467 222 24 less less ADJ iajs-3467 222 25 than than ADP iajs-3467 222 26 0.02 0.02 NUM iajs-3467 222 27 , , PUNCT iajs-3467 222 28 and and CCONJ iajs-3467 222 29 some some PRON iajs-3467 222 30 were be AUX iajs-3467 222 31 significant significant ADJ iajs-3467 222 32 where where SCONJ iajs-3467 222 33 p p NOUN iajs-3467 222 34 value value NOUN iajs-3467 222 35 < < X iajs-3467 222 36 0.05 0.05 NUM iajs-3467 222 37 . . PUNCT iajs-3467 223 1 in in ADP iajs-3467 223 2 general general ADJ iajs-3467 223 3 , , PUNCT iajs-3467 223 4 6 6 NUM iajs-3467 223 5 variables variable NOUN iajs-3467 223 6 are be AUX iajs-3467 223 7 selected select VERB iajs-3467 223 8 as as ADP iajs-3467 223 9 important important ADJ iajs-3467 223 10 variables variable NOUN iajs-3467 223 11 in in ADP iajs-3467 223 12 the the DET iajs-3467 223 13 dataset dataset NOUN iajs-3467 223 14 . . PUNCT iajs-3467 224 1 table table NOUN iajs-3467 224 2 3 3 NUM iajs-3467 224 3 . . PUNCT iajs-3467 224 4 estimated estimate VERB iajs-3467 224 5 coefficient coefficient NOUN iajs-3467 224 6 for for ADP iajs-3467 224 7 model model NOUN iajs-3467 224 8 using use VERB iajs-3467 224 9 heart heart NOUN iajs-3467 224 10 disease disease NOUN iajs-3467 224 11 dataset dataset VERB iajs-3467 224 12 . . PUNCT iajs-3467 225 1 variables variable NOUN iajs-3467 225 2 in in ADP iajs-3467 225 3 the the DET iajs-3467 225 4 equation equation NOUN iajs-3467 225 5 οΏ½ οΏ½ PROPN iajs-3467 225 6 Μ‚ Μ‚ PROPN iajs-3467 225 7 οΏ½ οΏ½ PROPN iajs-3467 225 8 s.e s.e PROPN iajs-3467 225 9 . . PROPN iajs-3467 225 10 wald wald PROPN iajs-3467 225 11 p p NOUN iajs-3467 225 12 - - PUNCT iajs-3467 225 13 value value NOUN iajs-3467 225 14 exp(b exp(b PROPN iajs-3467 225 15 ) ) PUNCT iajs-3467 225 16 constant constant ADJ iajs-3467 225 17 -5.225 -5.225 PUNCT iajs-3467 225 18 1.315 1.315 NUM iajs-3467 225 19 15.782 15.782 NUM iajs-3467 225 20 < < X iajs-3467 225 21 0.001 0.001 NUM iajs-3467 225 22 * * SYM iajs-3467 225 23 * * PROPN iajs-3467 225 24 0.005 0.005 NUM iajs-3467 225 25 sbp sbp NOUN iajs-3467 225 26 0.007 0.007 NUM iajs-3467 225 27 0.006 0.006 NUM iajs-3467 225 28 1.288 1.288 NUM iajs-3467 225 29 0.256 0.256 NUM iajs-3467 225 30 1.007 1.007 NUM iajs-3467 225 31 tobacco tobacco NOUN iajs-3467 225 32 0.379 0.379 NUM iajs-3467 225 33 0.027 0.027 NUM iajs-3467 225 34 8.903 8.903 NUM iajs-3467 225 35 0.003 0.003 NUM iajs-3467 225 36 * * SYM iajs-3467 225 37 1.083 1.083 NUM iajs-3467 225 38 ldl ldl NOUN iajs-3467 225 39 0.174 0.174 NUM iajs-3467 225 40 0.06 0.06 NUM iajs-3467 225 41 8.498 8.498 NUM iajs-3467 225 42 0.004 0.004 NUM iajs-3467 225 43 * * NUM iajs-3467 225 44 1.19 1.19 NUM iajs-3467 225 45 adiposity adiposity NOUN iajs-3467 225 46 0.019 0.019 NUM iajs-3467 225 47 0.029 0.029 NUM iajs-3467 225 48 0.403 0.403 NUM iajs-3467 225 49 0.526 0.526 NUM iajs-3467 225 50 1.019 1.019 NUM iajs-3467 225 51 famhist famhist PROPN iajs-3467 225 52 -0.925 -0.925 ADP iajs-3467 225 53 0.228 0.228 NUM iajs-3467 225 54 16.488 16.488 NUM iajs-3467 225 55 < < X iajs-3467 225 56 0.001 0.001 NUM iajs-3467 225 57 * * SYM iajs-3467 225 58 * * SYM iajs-3467 225 59 0.396 0.396 NUM iajs-3467 225 60 typea typea NOUN iajs-3467 225 61 0.340 0.340 NUM iajs-3467 225 62 0.012 0.012 NUM iajs-3467 225 63 10.329 10.329 NUM iajs-3467 225 64 0.001 0.001 NUM iajs-3467 225 65 * * NUM iajs-3467 225 66 1.04 1.04 NUM iajs-3467 225 67 obesity obesity NOUN iajs-3467 225 68 -0.063 -0.063 NUM iajs-3467 225 69 0.044 0.044 NUM iajs-3467 225 70 2.021 2.021 NUM iajs-3467 225 71 0.155 0.155 NUM iajs-3467 225 72 0.939 0.939 NUM iajs-3467 225 73 alcohol alcohol NOUN iajs-3467 225 74 0.002 0.002 NUM iajs-3467 225 75 0.004 0.004 NUM iajs-3467 225 76 0.001 0.001 NUM iajs-3467 225 77 0.978 0.978 NUM iajs-3467 225 78 1 1 NUM iajs-3467 225 79 age age NOUN iajs-3467 225 80 0.453 0.453 NUM iajs-3467 225 81 0.012 0.012 NUM iajs-3467 225 82 13.901 13.901 NUM iajs-3467 225 83 < < NOUN iajs-3467 225 84 0.001 0.001 NUM iajs-3467 225 85 * * NOUN iajs-3467 225 86 * * NOUN iajs-3467 225 87 1.046 1.046 NUM iajs-3467 225 88 * * PUNCT iajs-3467 225 89 significant significant ADJ iajs-3467 225 90 effect effect NOUN iajs-3467 225 91 for for ADP iajs-3467 225 92 the the DET iajs-3467 225 93 parameter parameter NOUN iajs-3467 225 94 in in ADP iajs-3467 225 95 the the DET iajs-3467 225 96 equation equation NOUN iajs-3467 225 97 . . PUNCT iajs-3467 226 1 * * PUNCT iajs-3467 226 2 * * PUNCT iajs-3467 226 3 high high ADJ iajs-3467 226 4 significant significant ADJ iajs-3467 226 5 effect effect NOUN iajs-3467 226 6 for for ADP iajs-3467 226 7 the the DET iajs-3467 226 8 parameter parameter NOUN iajs-3467 226 9 in in ADP iajs-3467 226 10 the the DET iajs-3467 226 11 equation equation NOUN iajs-3467 226 12 . . PUNCT iajs-3467 227 1 ihjpas ihjpas PROPN iajs-3467 227 2 . . PUNCT iajs-3467 228 1 37 37 NUM iajs-3467 228 2 ( ( PUNCT iajs-3467 228 3 1 1 NUM iajs-3467 228 4 ) ) PUNCT iajs-3467 228 5 2024 2024 NUM iajs-3467 228 6 423 423 NUM iajs-3467 228 7 instead instead ADV iajs-3467 228 8 of of ADP iajs-3467 228 9 using use VERB iajs-3467 228 10 6 6 NUM iajs-3467 228 11 variables variable NOUN iajs-3467 228 12 in in ADP iajs-3467 228 13 the the DET iajs-3467 228 14 model model NOUN iajs-3467 228 15 as as SCONJ iajs-3467 228 16 it it PRON iajs-3467 228 17 was be AUX iajs-3467 228 18 illustrated illustrate VERB iajs-3467 228 19 in in ADP iajs-3467 228 20 table table NOUN iajs-3467 228 21 ( ( PUNCT iajs-3467 228 22 3 3 NUM iajs-3467 228 23 ) ) PUNCT iajs-3467 228 24 , , PUNCT iajs-3467 228 25 the the DET iajs-3467 228 26 principal principal ADJ iajs-3467 228 27 component component NOUN iajs-3467 228 28 variable variable NOUN iajs-3467 228 29 can can AUX iajs-3467 228 30 be be AUX iajs-3467 228 31 used use VERB iajs-3467 228 32 to to PART iajs-3467 228 33 reduce reduce VERB iajs-3467 228 34 the the DET iajs-3467 228 35 components component NOUN iajs-3467 228 36 . . PUNCT iajs-3467 229 1 in in ADP iajs-3467 229 2 figure figure NOUN iajs-3467 229 3 6.a 6.a NUM iajs-3467 229 4 , , PUNCT iajs-3467 229 5 we we PRON iajs-3467 229 6 draw draw VERB iajs-3467 229 7 a a DET iajs-3467 229 8 curve curve NOUN iajs-3467 229 9 as as ADP iajs-3467 229 10 a a DET iajs-3467 229 11 scree scree ADJ iajs-3467 229 12 plot plot NOUN iajs-3467 229 13 to to PART iajs-3467 229 14 show show VERB iajs-3467 229 15 the the DET iajs-3467 229 16 number number NOUN iajs-3467 229 17 of of ADP iajs-3467 229 18 components component NOUN iajs-3467 229 19 depends depend VERB iajs-3467 229 20 on on ADP iajs-3467 229 21 their their PRON iajs-3467 229 22 eigenvalues eigenvalue NOUN iajs-3467 229 23 . . PUNCT iajs-3467 230 1 using use VERB iajs-3467 230 2 an an DET iajs-3467 230 3 eigenvalue eigenvalue NOUN iajs-3467 230 4 equal equal ADJ iajs-3467 230 5 to to ADP iajs-3467 230 6 1 1 NUM iajs-3467 230 7 as as ADP iajs-3467 230 8 a a DET iajs-3467 230 9 cut cut NOUN iajs-3467 230 10 off off ADP iajs-3467 230 11 , , PUNCT iajs-3467 230 12 three three NUM iajs-3467 230 13 components component NOUN iajs-3467 230 14 ( ( PUNCT iajs-3467 230 15 rc1 rc1 PROPN iajs-3467 230 16 , , PUNCT iajs-3467 230 17 rc2 rc2 PROPN iajs-3467 230 18 , , PUNCT iajs-3467 230 19 rc3 rc3 PROPN iajs-3467 230 20 ) ) PUNCT iajs-3467 230 21 are be AUX iajs-3467 230 22 satisfactory satisfactory ADJ iajs-3467 230 23 to to PART iajs-3467 230 24 represent represent VERB iajs-3467 230 25 all all DET iajs-3467 230 26 the the DET iajs-3467 230 27 variables variable NOUN iajs-3467 230 28 in in ADP iajs-3467 230 29 the the DET iajs-3467 230 30 dataset dataset NOUN iajs-3467 230 31 . . PUNCT iajs-3467 231 1 figure figure NOUN iajs-3467 231 2 6.b 6.b NUM iajs-3467 231 3 shows show VERB iajs-3467 231 4 the the DET iajs-3467 231 5 connection connection NOUN iajs-3467 231 6 between between ADP iajs-3467 231 7 the the DET iajs-3467 231 8 variables variable NOUN iajs-3467 231 9 and and CCONJ iajs-3467 231 10 their their PRON iajs-3467 231 11 components component NOUN iajs-3467 231 12 . . PUNCT iajs-3467 232 1 a. a. NOUN iajs-3467 232 2 number number NOUN iajs-3467 232 3 of of ADP iajs-3467 232 4 principles principle NOUN iajs-3467 232 5 that that PRON iajs-3467 232 6 should should AUX iajs-3467 232 7 be be AUX iajs-3467 232 8 used use VERB iajs-3467 232 9 in in ADP iajs-3467 232 10 edsgd edsgd VERB iajs-3467 232 11 svm svm PROPN iajs-3467 232 12 b. b. PROPN iajs-3467 232 13 diagram diagram PROPN iajs-3467 232 14 for for ADP iajs-3467 232 15 connection connection NOUN iajs-3467 232 16 between between ADP iajs-3467 232 17 variables variable NOUN iajs-3467 232 18 and and CCONJ iajs-3467 232 19 their their PRON iajs-3467 232 20 component component NOUN iajs-3467 232 21 . . PUNCT iajs-3467 233 1 figure figure NOUN iajs-3467 233 2 6 6 NUM iajs-3467 233 3 . . PUNCT iajs-3467 234 1 importance importance NOUN iajs-3467 234 2 variables variable NOUN iajs-3467 234 3 for for SCONJ iajs-3467 234 4 heart heart NOUN iajs-3467 234 5 disease disease NOUN iajs-3467 234 6 dataset dataset VERB iajs-3467 234 7 using use VERB iajs-3467 234 8 pca pca NOUN iajs-3467 234 9 technique technique NOUN iajs-3467 234 10 . . PUNCT iajs-3467 235 1 the the DET iajs-3467 235 2 summary summary NOUN iajs-3467 235 3 of of ADP iajs-3467 235 4 the the DET iajs-3467 235 5 main main ADJ iajs-3467 235 6 principles principle NOUN iajs-3467 235 7 and and CCONJ iajs-3467 235 8 their their PRON iajs-3467 235 9 weight weight NOUN iajs-3467 235 10 according accord VERB iajs-3467 235 11 to to ADP iajs-3467 235 12 the the DET iajs-3467 235 13 variable variable NOUN iajs-3467 235 14 that that PRON iajs-3467 235 15 is be AUX iajs-3467 235 16 used use VERB iajs-3467 235 17 in in ADP iajs-3467 235 18 the the DET iajs-3467 235 19 real real ADJ iajs-3467 235 20 dataset dataset NOUN iajs-3467 235 21 is be AUX iajs-3467 235 22 shown show VERB iajs-3467 235 23 in in ADP iajs-3467 235 24 table table NOUN iajs-3467 235 25 4 4 NUM iajs-3467 235 26 . . PUNCT iajs-3467 235 27 table table NOUN iajs-3467 235 28 4 4 NUM iajs-3467 235 29 . . PUNCT iajs-3467 235 30 principle principle ADJ iajs-3467 235 31 components component NOUN iajs-3467 235 32 for for ADP iajs-3467 235 33 the the DET iajs-3467 235 34 dataset dataset NOUN iajs-3467 235 35 . . PUNCT iajs-3467 236 1 component component NOUN iajs-3467 236 2 loadings loading NOUN iajs-3467 236 3 variables variable NOUN iajs-3467 236 4 rc1 rc1 PROPN iajs-3467 236 5 rc2 rc2 PROPN iajs-3467 236 6 rc3 rc3 PROPN iajs-3467 236 7 sbp sbp PROPN iajs-3467 236 8 tobacco tobacco NOUN iajs-3467 236 9 0.683 0.683 NUM iajs-3467 236 10 ldl ldl NOUN iajs-3467 236 11 0.747 0.747 NUM iajs-3467 236 12 adiposity adiposity NOUN iajs-3467 236 13 0.868 0.868 NUM iajs-3467 236 14 typea typea NOUN iajs-3467 236 15 0.959 0.959 NUM iajs-3467 236 16 obesity obesity NOUN iajs-3467 236 17 0.877 0.877 NUM iajs-3467 236 18 alcohol alcohol NOUN iajs-3467 236 19 0.853 0.853 NUM iajs-3467 236 20 age age NOUN iajs-3467 236 21 0.477 0.477 NUM iajs-3467 236 22 figure figure NOUN iajs-3467 236 23 7 7 NUM iajs-3467 236 24 shows show VERB iajs-3467 236 25 the the DET iajs-3467 236 26 important important ADJ iajs-3467 236 27 variables variable NOUN iajs-3467 236 28 and and CCONJ iajs-3467 236 29 their their PRON iajs-3467 236 30 connections connection NOUN iajs-3467 236 31 . . PUNCT iajs-3467 237 1 in in ADP iajs-3467 237 2 figure figure NOUN iajs-3467 237 3 7.a 7.a NOUN iajs-3467 237 4 the the DET iajs-3467 237 5 important important ADJ iajs-3467 237 6 variables variable NOUN iajs-3467 237 7 are be AUX iajs-3467 237 8 sorted sort VERB iajs-3467 237 9 in in ADP iajs-3467 237 10 order order NOUN iajs-3467 237 11 . . PUNCT iajs-3467 238 1 it it PRON iajs-3467 238 2 shows show VERB iajs-3467 238 3 that that SCONJ iajs-3467 238 4 age age NOUN iajs-3467 238 5 and and CCONJ iajs-3467 238 6 ldl ldl NOUN iajs-3467 238 7 are be AUX iajs-3467 238 8 the the DET iajs-3467 238 9 most most ADV iajs-3467 238 10 important important ADJ iajs-3467 238 11 variables variable NOUN iajs-3467 238 12 followed follow VERB iajs-3467 238 13 by by ADP iajs-3467 238 14 tobacco tobacco NOUN iajs-3467 238 15 and and CCONJ iajs-3467 238 16 typea typea NOUN iajs-3467 238 17 . . PUNCT iajs-3467 239 1 figure figure NOUN iajs-3467 239 2 7.b 7.b NUM iajs-3467 239 3 shows show VERB iajs-3467 239 4 the the DET iajs-3467 239 5 connection connection NOUN iajs-3467 239 6 between between ADP iajs-3467 239 7 variables variable NOUN iajs-3467 239 8 . . PUNCT iajs-3467 240 1 we we PRON iajs-3467 240 2 see see VERB iajs-3467 240 3 the the DET iajs-3467 240 4 strong strong ADJ iajs-3467 240 5 connections connection NOUN iajs-3467 240 6 are be AUX iajs-3467 240 7 marked mark VERB iajs-3467 240 8 with with ADP iajs-3467 240 9 thick thick ADJ iajs-3467 240 10 blue blue ADJ iajs-3467 240 11 lines line NOUN iajs-3467 240 12 , , PUNCT iajs-3467 240 13 and and CCONJ iajs-3467 240 14 the the DET iajs-3467 240 15 weak weak ADJ iajs-3467 240 16 connection connection NOUN iajs-3467 240 17 were be AUX iajs-3467 240 18 marked mark VERB iajs-3467 240 19 with with ADP iajs-3467 240 20 thin thin ADJ iajs-3467 240 21 blue blue ADJ iajs-3467 240 22 lines line NOUN iajs-3467 240 23 . . PUNCT iajs-3467 241 1 ihjpas ihjpas PROPN iajs-3467 241 2 . . PUNCT iajs-3467 242 1 37 37 NUM iajs-3467 242 2 ( ( PUNCT iajs-3467 242 3 1 1 NUM iajs-3467 242 4 ) ) PUNCT iajs-3467 242 5 2024 2024 NUM iajs-3467 242 6 424 424 NUM iajs-3467 242 7 a. a. NOUN iajs-3467 242 8 sorting sort VERB iajs-3467 242 9 variables variable NOUN iajs-3467 242 10 depend depend VERB iajs-3467 242 11 on on ADP iajs-3467 242 12 their their PRON iajs-3467 242 13 importance importance NOUN iajs-3467 242 14 b. b. NOUN iajs-3467 243 1 the the DET iajs-3467 243 2 connection connection NOUN iajs-3467 243 3 between between ADP iajs-3467 243 4 variables variable NOUN iajs-3467 243 5 figure figure VERB iajs-3467 243 6 7 7 NUM iajs-3467 243 7 . . NOUN iajs-3467 243 8 path path NOUN iajs-3467 243 9 diagram diagram NOUN iajs-3467 243 10 for for ADP iajs-3467 243 11 the the DET iajs-3467 243 12 relation relation NOUN iajs-3467 243 13 between between ADP iajs-3467 243 14 the the DET iajs-3467 243 15 variables variable NOUN iajs-3467 243 16 and and CCONJ iajs-3467 243 17 the the DET iajs-3467 243 18 component component NOUN iajs-3467 243 19 . . PUNCT iajs-3467 244 1 8 8 X iajs-3467 244 2 . . X iajs-3467 244 3 results result NOUN iajs-3467 244 4 and and CCONJ iajs-3467 244 5 discussion discussion VERB iajs-3467 244 6 the the DET iajs-3467 244 7 enhanced enhanced ADJ iajs-3467 244 8 version version NOUN iajs-3467 244 9 of of ADP iajs-3467 244 10 svm svm PROPN iajs-3467 244 11 ( ( PUNCT iajs-3467 244 12 esgd esgd NOUN iajs-3467 244 13 - - PUNCT iajs-3467 244 14 svm svm NOUN iajs-3467 244 15 ) ) PUNCT iajs-3467 244 16 is be AUX iajs-3467 244 17 applied apply VERB iajs-3467 244 18 to to ADP iajs-3467 244 19 the the DET iajs-3467 244 20 heart heart NOUN iajs-3467 244 21 disease disease NOUN iajs-3467 244 22 dataset dataset VERB iajs-3467 244 23 . . PUNCT iajs-3467 245 1 a a DET iajs-3467 245 2 comparison comparison NOUN iajs-3467 245 3 between between ADP iajs-3467 245 4 some some DET iajs-3467 245 5 different different ADJ iajs-3467 245 6 kernel kernel NOUN iajs-3467 245 7 functions function NOUN iajs-3467 245 8 had have AUX iajs-3467 245 9 been be AUX iajs-3467 245 10 shown show VERB iajs-3467 245 11 . . PUNCT iajs-3467 246 1 in in ADP iajs-3467 246 2 application application NOUN iajs-3467 246 3 , , PUNCT iajs-3467 246 4 we we PRON iajs-3467 246 5 applied apply VERB iajs-3467 246 6 the the DET iajs-3467 246 7 most most ADV iajs-3467 246 8 common common ADJ iajs-3467 246 9 kernel kernel NOUN iajs-3467 246 10 which which PRON iajs-3467 246 11 is be AUX iajs-3467 246 12 linear linear ADJ iajs-3467 246 13 , , PUNCT iajs-3467 246 14 polynomial polynomial ADJ iajs-3467 246 15 , , PUNCT iajs-3467 246 16 rbf rbf PROPN iajs-3467 246 17 and and CCONJ iajs-3467 246 18 mlp mlp PROPN iajs-3467 246 19 kernels kernel NOUN iajs-3467 246 20 . . PUNCT iajs-3467 247 1 in in ADP iajs-3467 247 2 addition addition NOUN iajs-3467 247 3 , , PUNCT iajs-3467 247 4 the the DET iajs-3467 247 5 proposed propose VERB iajs-3467 247 6 method method NOUN iajs-3467 247 7 was be AUX iajs-3467 247 8 compared compare VERB iajs-3467 247 9 with with ADP iajs-3467 247 10 some some DET iajs-3467 247 11 common common ADJ iajs-3467 247 12 methods method NOUN iajs-3467 247 13 . . PUNCT iajs-3467 248 1 these these DET iajs-3467 248 2 methods method NOUN iajs-3467 248 3 used use VERB iajs-3467 248 4 the the DET iajs-3467 248 5 same same ADJ iajs-3467 248 6 dataset dataset NOUN iajs-3467 248 7 for for ADP iajs-3467 248 8 leukemia leukemia NOUN iajs-3467 248 9 classification classification NOUN iajs-3467 248 10 which which PRON iajs-3467 248 11 are be AUX iajs-3467 248 12 k k ADJ iajs-3467 248 13 - - PUNCT iajs-3467 248 14 nearest near ADJ iajs-3467 248 15 neighbor neighbor NOUN iajs-3467 248 16 random random ADJ iajs-3467 248 17 forest forest NOUN iajs-3467 248 18 and and CCONJ iajs-3467 248 19 naΓ―ve naΓ―ve ADJ iajs-3467 248 20 bayes bayes NOUN iajs-3467 248 21 . . PUNCT iajs-3467 249 1 in in ADP iajs-3467 249 2 figure10 figure10 PROPN iajs-3467 249 3 , , PUNCT iajs-3467 249 4 esgd esgd NOUN iajs-3467 249 5 - - PUNCT iajs-3467 249 6 svm svm NOUN iajs-3467 249 7 classification classification NOUN iajs-3467 249 8 method method NOUN iajs-3467 249 9 for for ADP iajs-3467 249 10 the the DET iajs-3467 249 11 heart heart NOUN iajs-3467 249 12 disease disease NOUN iajs-3467 249 13 dataset dataset NOUN iajs-3467 249 14 is be AUX iajs-3467 249 15 plotted plot VERB iajs-3467 249 16 . . PUNCT iajs-3467 250 1 the the DET iajs-3467 250 2 two two NUM iajs-3467 250 3 most most ADV iajs-3467 250 4 important important ADJ iajs-3467 250 5 variables variable NOUN iajs-3467 250 6 , , PUNCT iajs-3467 250 7 which which PRON iajs-3467 250 8 are be AUX iajs-3467 250 9 tobacco tobacco NOUN iajs-3467 250 10 and and CCONJ iajs-3467 250 11 ldl ldl NOUN iajs-3467 250 12 are be AUX iajs-3467 250 13 used use VERB iajs-3467 250 14 for for ADP iajs-3467 250 15 visualization visualization NOUN iajs-3467 250 16 . . PUNCT iajs-3467 251 1 as as SCONJ iajs-3467 251 2 it it PRON iajs-3467 251 3 is be AUX iajs-3467 251 4 shown show VERB iajs-3467 251 5 from from ADP iajs-3467 251 6 the the DET iajs-3467 251 7 plot plot NOUN iajs-3467 251 8 ( ( PUNCT iajs-3467 251 9 figure figure NOUN iajs-3467 251 10 8) 8) NUM iajs-3467 251 11 and and CCONJ iajs-3467 251 12 the the DET iajs-3467 251 13 table table NOUN iajs-3467 251 14 ( ( PUNCT iajs-3467 251 15 table table NOUN iajs-3467 251 16 5 5 NUM iajs-3467 251 17 ) ) PUNCT iajs-3467 251 18 the the DET iajs-3467 251 19 best good ADJ iajs-3467 251 20 version version NOUN iajs-3467 251 21 for for ADP iajs-3467 251 22 esgdsvm esgdsvm NOUN iajs-3467 251 23 method method NOUN iajs-3467 251 24 is be AUX iajs-3467 251 25 satisfied satisfied ADJ iajs-3467 251 26 when when SCONJ iajs-3467 251 27 the the DET iajs-3467 251 28 rbf rbf PROPN iajs-3467 251 29 kernel kernel PROPN iajs-3467 251 30 is be AUX iajs-3467 251 31 applied apply VERB iajs-3467 251 32 . . PUNCT iajs-3467 252 1 the the DET iajs-3467 252 2 mlp mlp PROPN iajs-3467 252 3 kernel kernel PROPN iajs-3467 252 4 gets get VERB iajs-3467 252 5 98.10 98.10 NUM iajs-3467 252 6 % % NOUN iajs-3467 252 7 accuracy accuracy NOUN iajs-3467 252 8 which which PRON iajs-3467 252 9 is be AUX iajs-3467 252 10 the the DET iajs-3467 252 11 highest high ADJ iajs-3467 252 12 performance performance NOUN iajs-3467 252 13 compared compare VERB iajs-3467 252 14 with with ADP iajs-3467 252 15 other other ADJ iajs-3467 252 16 kernels kernel NOUN iajs-3467 252 17 . . PUNCT iajs-3467 253 1 sgd sgd NOUN iajs-3467 253 2 - - PUNCT iajs-3467 253 3 svm svm PROPN iajs-3467 253 4 performed perform VERB iajs-3467 253 5 96.53 96.53 NUM iajs-3467 253 6 % % NOUN iajs-3467 253 7 accuracy accuracy NOUN iajs-3467 253 8 for for ADP iajs-3467 253 9 the the DET iajs-3467 253 10 linear linear ADJ iajs-3467 253 11 kernel kernel NOUN iajs-3467 253 12 , , PUNCT iajs-3467 253 13 98.03 98.03 NUM iajs-3467 253 14 % % NOUN iajs-3467 253 15 accuracy accuracy NOUN iajs-3467 253 16 for for ADP iajs-3467 253 17 the the DET iajs-3467 253 18 polynomial polynomial ADJ iajs-3467 253 19 kernel kernel NOUN iajs-3467 253 20 , , PUNCT iajs-3467 253 21 and and CCONJ iajs-3467 253 22 96.53 96.53 NUM iajs-3467 253 23 % % NOUN iajs-3467 253 24 accuracy accuracy NOUN iajs-3467 253 25 for for ADP iajs-3467 253 26 rbf rbf PROPN iajs-3467 253 27 kernel kernel PROPN iajs-3467 253 28 . . PUNCT iajs-3467 254 1 esgd esgd PROPN iajs-3467 254 2 - - PUNCT iajs-3467 254 3 svm svm PROPN iajs-3467 254 4 performed perform VERB iajs-3467 254 5 much much ADV iajs-3467 254 6 better well ADJ iajs-3467 254 7 than than ADP iajs-3467 254 8 other other ADJ iajs-3467 254 9 methods method NOUN iajs-3467 254 10 where where SCONJ iajs-3467 254 11 logistic logistic ADJ iajs-3467 254 12 regression regression NOUN iajs-3467 254 13 performed perform VERB iajs-3467 254 14 87.42 87.42 NUM iajs-3467 254 15 % % NOUN iajs-3467 254 16 accuracy accuracy NOUN iajs-3467 254 17 , , PUNCT iajs-3467 254 18 k k X iajs-3467 254 19 - - PUNCT iajs-3467 254 20 nearest near ADJ iajs-3467 254 21 neighbor neighbor NOUN iajs-3467 254 22 performed perform VERB iajs-3467 254 23 86.70 86.70 NUM iajs-3467 254 24 % % NOUN iajs-3467 254 25 accuracy accuracy NOUN iajs-3467 254 26 , , PUNCT iajs-3467 254 27 random random ADJ iajs-3467 254 28 forest forest NOUN iajs-3467 254 29 performed perform VERB iajs-3467 254 30 87.33 87.33 NUM iajs-3467 254 31 % % NOUN iajs-3467 254 32 accuracy accuracy NOUN iajs-3467 254 33 , , PUNCT iajs-3467 254 34 and and CCONJ iajs-3467 254 35 naΓ―ve naΓ―ve ADJ iajs-3467 254 36 bayes bayes NOUN iajs-3467 254 37 performed perform VERB iajs-3467 254 38 84.34 84.34 NUM iajs-3467 254 39 % % NOUN iajs-3467 254 40 accuracy accuracy NOUN iajs-3467 254 41 . . PUNCT iajs-3467 255 1 a. a. PROPN iajs-3467 255 2 linear linear PROPN iajs-3467 255 3 kernel kernel PROPN iajs-3467 255 4 b. b. PROPN iajs-3467 255 5 polynomial polynomial PROPN iajs-3467 255 6 kernel kernel PROPN iajs-3467 255 7 ihjpas ihjpas PROPN iajs-3467 255 8 . . PUNCT iajs-3467 256 1 37 37 NUM iajs-3467 256 2 ( ( PUNCT iajs-3467 256 3 1 1 NUM iajs-3467 256 4 ) ) PUNCT iajs-3467 256 5 2024 2024 NUM iajs-3467 256 6 425 425 NUM iajs-3467 256 7 c. c. PROPN iajs-3467 256 8 rbf rbf PROPN iajs-3467 256 9 kernel kernel PROPN iajs-3467 256 10 d. d. PROPN iajs-3467 256 11 mlp mlp PROPN iajs-3467 256 12 kernel kernel PROPN iajs-3467 256 13 figure figure VERB iajs-3467 256 14 8 8 NUM iajs-3467 256 15 . . PUNCT iajs-3467 257 1 classification classification NOUN iajs-3467 257 2 for for ADP iajs-3467 257 3 heart heart NOUN iajs-3467 257 4 disease disease NOUN iajs-3467 257 5 dataset dataset VERB iajs-3467 257 6 using use VERB iajs-3467 257 7 esgd esgd NOUN iajs-3467 257 8 - - PUNCT iajs-3467 257 9 svm svm PROPN iajs-3467 257 10 . . NOUN iajs-3467 257 11 table table NOUN iajs-3467 257 12 5 5 NUM iajs-3467 257 13 . . PUNCT iajs-3467 257 14 comparison comparison NOUN iajs-3467 257 15 between between ADP iajs-3467 257 16 esgd esgd NOUN iajs-3467 257 17 - - PUNCT iajs-3467 257 18 svm svm ADJ iajs-3467 257 19 , , PUNCT iajs-3467 257 20 logistic logistic ADJ iajs-3467 257 21 regression regression NOUN iajs-3467 257 22 , , PUNCT iajs-3467 257 23 k k X iajs-3467 257 24 - - PUNCT iajs-3467 257 25 nearest near ADJ iajs-3467 257 26 neighbors neighbor NOUN iajs-3467 257 27 , , PUNCT iajs-3467 257 28 random random ADJ iajs-3467 257 29 forest forest NOUN iajs-3467 257 30 , , PUNCT iajs-3467 257 31 and and CCONJ iajs-3467 257 32 naive naive ADJ iajs-3467 257 33 bayes bayes NOUN iajs-3467 257 34 for for ADP iajs-3467 257 35 classification classification NOUN iajs-3467 257 36 heart heart NOUN iajs-3467 257 37 disease disease NOUN iajs-3467 257 38 dataset dataset NOUN iajs-3467 257 39 . . PUNCT iajs-3467 258 1 methods method NOUN iajs-3467 258 2 number number NOUN iajs-3467 258 3 of of ADP iajs-3467 258 4 support support NOUN iajs-3467 258 5 vectors vector NOUN iajs-3467 258 6 sensitivity sensitivity NOUN iajs-3467 258 7 % % NOUN iajs-3467 258 8 specificity specificity NOUN iajs-3467 258 9 % % NOUN iajs-3467 258 10 accuracy accuracy NOUN iajs-3467 258 11 rate rate NOUN iajs-3467 258 12 % % NOUN iajs-3467 258 13 esgd esgd NOUN iajs-3467 258 14 - - PUNCT iajs-3467 258 15 svm svm PROPN iajs-3467 258 16 linear linear PROPN iajs-3467 258 17 kernel kernel PROPN iajs-3467 258 18 42 42 NUM iajs-3467 258 19 96.61 96.61 NUM iajs-3467 258 20 98.67 98.67 NUM iajs-3467 258 21 96.53 96.53 NUM iajs-3467 258 22 polynomial polynomial ADJ iajs-3467 258 23 kernel kernel NOUN iajs-3467 259 1 96 96 NUM iajs-3467 259 2 98.19 98.19 NUM iajs-3467 259 3 94.19 94.19 NUM iajs-3467 259 4 98.03 98.03 NUM iajs-3467 259 5 rbf rbf PROPN iajs-3467 259 6 kernel kernel PROPN iajs-3467 259 7 51 51 NUM iajs-3467 259 8 96.61 96.61 NUM iajs-3467 259 9 98.67 98.67 NUM iajs-3467 259 10 96.53 96.53 NUM iajs-3467 259 11 mlp mlp PROPN iajs-3467 259 12 kernel kernel PROPN iajs-3467 259 13 48 48 NUM iajs-3467 259 14 97.21 97.21 NUM iajs-3467 259 15 99.10 99.10 NUM iajs-3467 259 16 98.10 98.10 NUM iajs-3467 259 17 logistic logistic ADJ iajs-3467 259 18 regression regression NOUN iajs-3467 259 19 86.39 86.39 NUM iajs-3467 259 20 89.48 89.48 NUM iajs-3467 259 21 87.42 87.42 NUM iajs-3467 259 22 k k X iajs-3467 259 23 - - PUNCT iajs-3467 259 24 nearest near ADJ iajs-3467 259 25 neighbors neighbor NOUN iajs-3467 259 26 85.20 85.20 NUM iajs-3467 259 27 88.40 88.40 NUM iajs-3467 259 28 86.70 86.70 NUM iajs-3467 259 29 random random ADJ iajs-3467 259 30 forest forest NOUN iajs-3467 259 31 89.33 89.33 NUM iajs-3467 259 32 84.87 84.87 NUM iajs-3467 259 33 87.33 87.33 NUM iajs-3467 259 34 naive naive ADJ iajs-3467 259 35 bayes baye NOUN iajs-3467 259 36 86.49 86.49 NUM iajs-3467 259 37 82.71 82.71 NUM iajs-3467 259 38 84.34 84.34 NUM iajs-3467 259 39 the the DET iajs-3467 259 40 receiver receiver ADV iajs-3467 259 41 operating operate VERB iajs-3467 259 42 characteristic characteristic ADJ iajs-3467 259 43 curve curve NOUN iajs-3467 259 44 , , PUNCT iajs-3467 259 45 or or CCONJ iajs-3467 259 46 roc roc PROPN iajs-3467 259 47 curve curve NOUN iajs-3467 259 48 , , PUNCT iajs-3467 259 49 is be AUX iajs-3467 259 50 shown show VERB iajs-3467 259 51 . . PUNCT iajs-3467 260 1 roc roc PROPN iajs-3467 260 2 is be AUX iajs-3467 260 3 a a DET iajs-3467 260 4 graphical graphical ADJ iajs-3467 260 5 plot plot NOUN iajs-3467 260 6 that that PRON iajs-3467 260 7 illustrates illustrate VERB iajs-3467 260 8 the the DET iajs-3467 260 9 diagnostic diagnostic ADJ iajs-3467 260 10 ability ability NOUN iajs-3467 260 11 of of ADP iajs-3467 260 12 a a DET iajs-3467 260 13 binary binary ADJ iajs-3467 260 14 classifier classifier NOUN iajs-3467 260 15 system system NOUN iajs-3467 260 16 as as SCONJ iajs-3467 260 17 its its PRON iajs-3467 260 18 discrimination discrimination NOUN iajs-3467 260 19 threshold threshold NOUN iajs-3467 260 20 is be AUX iajs-3467 260 21 varied varied ADJ iajs-3467 260 22 for for ADP iajs-3467 260 23 both both CCONJ iajs-3467 260 24 the the DET iajs-3467 260 25 logistic logistic ADJ iajs-3467 260 26 regression regression NOUN iajs-3467 260 27 model model NOUN iajs-3467 260 28 and and CCONJ iajs-3467 260 29 esgd esgd NOUN iajs-3467 260 30 - - PUNCT iajs-3467 260 31 svm svm NOUN iajs-3467 260 32 models model NOUN iajs-3467 260 33 . . PUNCT iajs-3467 261 1 a. a. PROPN iajs-3467 261 2 roc roc PROPN iajs-3467 261 3 for for ADP iajs-3467 261 4 esgd esgd NOUN iajs-3467 261 5 - - PUNCT iajs-3467 261 6 svm svm PROPN iajs-3467 261 7 model model NOUN iajs-3467 261 8 b. b. PROPN iajs-3467 261 9 roc roc PROPN iajs-3467 261 10 for for ADP iajs-3467 261 11 logistic logistic ADJ iajs-3467 261 12 regression regression NOUN iajs-3467 261 13 model model NOUN iajs-3467 261 14 figure figure NOUN iajs-3467 261 15 9 9 NUM iajs-3467 261 16 . . PUNCT iajs-3467 262 1 roc roc PROPN iajs-3467 262 2 curve curve NOUN iajs-3467 262 3 for for ADP iajs-3467 262 4 both both DET iajs-3467 262 5 esgd esgd NOUN iajs-3467 262 6 - - PUNCT iajs-3467 262 7 svm svm ADJ iajs-3467 262 8 and and CCONJ iajs-3467 262 9 logistic logistic ADJ iajs-3467 262 10 regression regression NOUN iajs-3467 262 11 models model NOUN iajs-3467 262 12 . . PUNCT iajs-3467 263 1 ihjpas ihjpas PROPN iajs-3467 263 2 . . PUNCT iajs-3467 264 1 37 37 NUM iajs-3467 264 2 ( ( PUNCT iajs-3467 264 3 1 1 NUM iajs-3467 264 4 ) ) PUNCT iajs-3467 264 5 2024 2024 NUM iajs-3467 264 6 426 426 NUM iajs-3467 264 7 9 9 NUM iajs-3467 264 8 . . PUNCT iajs-3467 264 9 conclusion conclusion NOUN iajs-3467 264 10 this this DET iajs-3467 264 11 paper paper NOUN iajs-3467 264 12 presents present VERB iajs-3467 264 13 the the DET iajs-3467 264 14 esgd esgd NOUN iajs-3467 264 15 - - PUNCT iajs-3467 264 16 svm svm PROPN iajs-3467 264 17 method method NOUN iajs-3467 264 18 . . PUNCT iajs-3467 265 1 the the DET iajs-3467 265 2 stochastic stochastic ADJ iajs-3467 265 3 gradient gradient ADJ iajs-3467 265 4 descent descent NOUN iajs-3467 265 5 process process NOUN iajs-3467 265 6 is be AUX iajs-3467 265 7 used use VERB iajs-3467 265 8 to to PART iajs-3467 265 9 develop develop VERB iajs-3467 265 10 the the DET iajs-3467 265 11 method method NOUN iajs-3467 265 12 . . PUNCT iajs-3467 266 1 the the DET iajs-3467 266 2 kernel kernel PROPN iajs-3467 266 3 transformation transformation NOUN iajs-3467 266 4 technique technique NOUN iajs-3467 266 5 and and CCONJ iajs-3467 266 6 dimensionality dimensionality NOUN iajs-3467 266 7 reduction reduction NOUN iajs-3467 266 8 for for ADP iajs-3467 266 9 variables variable NOUN iajs-3467 266 10 are be AUX iajs-3467 266 11 used use VERB iajs-3467 266 12 to to PART iajs-3467 266 13 achieve achieve VERB iajs-3467 266 14 the the DET iajs-3467 266 15 best good ADJ iajs-3467 266 16 classification classification NOUN iajs-3467 266 17 accuracy accuracy NOUN iajs-3467 266 18 with with ADP iajs-3467 266 19 esgd esgd NOUN iajs-3467 266 20 - - PUNCT iajs-3467 266 21 svm svm NOUN iajs-3467 266 22 . . PROPN iajs-3467 267 1 two two NUM iajs-3467 267 2 simulation simulation NOUN iajs-3467 267 3 datasets dataset NOUN iajs-3467 267 4 are be AUX iajs-3467 267 5 used use VERB iajs-3467 267 6 to to PART iajs-3467 267 7 test test VERB iajs-3467 267 8 the the DET iajs-3467 267 9 implementation implementation NOUN iajs-3467 267 10 of of ADP iajs-3467 267 11 the the DET iajs-3467 267 12 method method NOUN iajs-3467 267 13 . . PUNCT iajs-3467 268 1 the the DET iajs-3467 268 2 results result NOUN iajs-3467 268 3 show show VERB iajs-3467 268 4 that that SCONJ iajs-3467 268 5 esgd esgd NOUN iajs-3467 268 6 - - PUNCT iajs-3467 268 7 svm svm PROPN iajs-3467 268 8 has have VERB iajs-3467 268 9 higher high ADJ iajs-3467 268 10 accuracy accuracy NOUN iajs-3467 268 11 compared compare VERB iajs-3467 268 12 to to ADP iajs-3467 268 13 some some DET iajs-3467 268 14 other other ADJ iajs-3467 268 15 classification classification NOUN iajs-3467 268 16 methods method NOUN iajs-3467 268 17 : : PUNCT iajs-3467 268 18 logistic logistic ADJ iajs-3467 268 19 regression regression NOUN iajs-3467 268 20 , , PUNCT iajs-3467 268 21 k k X iajs-3467 268 22 - - PUNCT iajs-3467 268 23 nearest near ADJ iajs-3467 268 24 neighbors neighbor NOUN iajs-3467 268 25 , , PUNCT iajs-3467 268 26 random random ADJ iajs-3467 268 27 forest forest NOUN iajs-3467 268 28 and and CCONJ iajs-3467 268 29 naive naive ADJ iajs-3467 268 30 bayes baye NOUN iajs-3467 268 31 . . PUNCT iajs-3467 269 1 when when SCONJ iajs-3467 269 2 the the DET iajs-3467 269 3 method method NOUN iajs-3467 269 4 was be AUX iajs-3467 269 5 applied apply VERB iajs-3467 269 6 to to ADP iajs-3467 269 7 a a DET iajs-3467 269 8 real real ADJ iajs-3467 269 9 dataset dataset NOUN iajs-3467 269 10 ( ( PUNCT iajs-3467 269 11 heart heart NOUN iajs-3467 269 12 disease disease NOUN iajs-3467 269 13 ) ) PUNCT iajs-3467 269 14 , , PUNCT iajs-3467 269 15 it it PRON iajs-3467 269 16 was be AUX iajs-3467 269 17 found find VERB iajs-3467 269 18 that that SCONJ iajs-3467 269 19 the the DET iajs-3467 269 20 highest high ADJ iajs-3467 269 21 accuracy accuracy NOUN iajs-3467 269 22 ( ( PUNCT iajs-3467 269 23 98.10 98.10 NUM iajs-3467 269 24 % % NOUN iajs-3467 269 25 ) ) PUNCT iajs-3467 269 26 was be AUX iajs-3467 269 27 achieved achieve VERB iajs-3467 269 28 by by ADP iajs-3467 269 29 applying apply VERB iajs-3467 269 30 the the DET iajs-3467 269 31 mlp mlp NOUN iajs-3467 269 32 kernel kernel PROPN iajs-3467 269 33 . . PUNCT iajs-3467 270 1 acknowledgments acknowledgment NOUN iajs-3467 270 2 the the DET iajs-3467 270 3 authors author NOUN iajs-3467 270 4 are be AUX iajs-3467 270 5 very very ADV iajs-3467 270 6 grateful grateful ADJ iajs-3467 270 7 to to ADP iajs-3467 270 8 the the DET iajs-3467 270 9 reviewers reviewer NOUN iajs-3467 270 10 for for ADP iajs-3467 270 11 their their PRON iajs-3467 270 12 valuable valuable ADJ iajs-3467 270 13 comments comment NOUN iajs-3467 270 14 and and CCONJ iajs-3467 270 15 suggestions suggestion NOUN iajs-3467 270 16 to to PART iajs-3467 270 17 improve improve VERB iajs-3467 270 18 the the DET iajs-3467 270 19 paper paper NOUN iajs-3467 270 20 conflict conflict NOUN iajs-3467 270 21 of of ADP iajs-3467 270 22 interest interest NOUN iajs-3467 270 23 the the DET iajs-3467 270 24 authors author NOUN iajs-3467 270 25 declare declare VERB iajs-3467 270 26 that that SCONJ iajs-3467 270 27 they they PRON iajs-3467 270 28 have have VERB iajs-3467 270 29 no no DET iajs-3467 270 30 conflicts conflict NOUN iajs-3467 270 31 of of ADP iajs-3467 270 32 interest interest NOUN iajs-3467 270 33 . . PUNCT iajs-3467 271 1 funding funding NOUN iajs-3467 271 2 there there PRON iajs-3467 271 3 is be VERB iajs-3467 271 4 no no DET iajs-3467 271 5 financial financial ADJ iajs-3467 271 6 support support NOUN iajs-3467 271 7 in in ADP iajs-3467 271 8 preparation preparation NOUN iajs-3467 271 9 for for ADP iajs-3467 271 10 the the DET iajs-3467 271 11 publication publication NOUN iajs-3467 271 12 . . PUNCT iajs-3467 272 1 references reference NOUN iajs-3467 272 2 1 1 NUM iajs-3467 272 3 . . PUNCT iajs-3467 273 1 zou zou PROPN iajs-3467 273 2 , , PUNCT iajs-3467 273 3 x. x. PROPN iajs-3467 273 4 ; ; PUNCT iajs-3467 273 5 hu hu PROPN iajs-3467 273 6 , , PUNCT iajs-3467 273 7 y. y. PROPN iajs-3467 273 8 ; ; PUNCT iajs-3467 273 9 tian tian PROPN iajs-3467 273 10 , , PUNCT iajs-3467 273 11 z. z. PROPN iajs-3467 273 12 ; ; PUNCT iajs-3467 273 13 shen shen PROPN iajs-3467 273 14 , , PUNCT iajs-3467 273 15 k. k. PROPN iajs-3467 273 16 logistic logistic PROPN iajs-3467 273 17 regression regression NOUN iajs-3467 273 18 model model NOUN iajs-3467 273 19 optimization optimization NOUN iajs-3467 273 20 and and CCONJ iajs-3467 273 21 case case NOUN iajs-3467 273 22 analysis analysis NOUN iajs-3467 273 23 . . PUNCT iajs-3467 274 1 ieee ieee PROPN iajs-3467 274 2 7th 7th ADJ iajs-3467 274 3 international international ADJ iajs-3467 274 4 conference conference NOUN iajs-3467 274 5 on on ADP iajs-3467 274 6 computer computer NOUN iajs-3467 274 7 science science NOUN iajs-3467 274 8 and and CCONJ iajs-3467 274 9 network network NOUN iajs-3467 274 10 technology technology NOUN iajs-3467 274 11 ( ( PUNCT iajs-3467 274 12 iccsnt iccsnt PROPN iajs-3467 274 13 ) ) PUNCT iajs-3467 274 14 2019 2019 NUM iajs-3467 274 15 , , PUNCT iajs-3467 274 16 7 7 NUM iajs-3467 274 17 , , PUNCT iajs-3467 274 18 135 135 NUM iajs-3467 274 19 - - SYM iajs-3467 274 20 139 139 NUM iajs-3467 274 21 . . PUNCT iajs-3467 275 1 2 2 X iajs-3467 275 2 . . NUM iajs-3467 275 3 liaw liaw PROPN iajs-3467 275 4 , , PUNCT iajs-3467 275 5 a. a. NOUN iajs-3467 275 6 ; ; PUNCT iajs-3467 275 7 wiener wiener NOUN iajs-3467 275 8 , , PUNCT iajs-3467 275 9 m. m. NOUN iajs-3467 275 10 classification classification NOUN iajs-3467 275 11 and and CCONJ iajs-3467 275 12 regression regression NOUN iajs-3467 275 13 by by ADP iajs-3467 275 14 random random ADJ iajs-3467 275 15 forest forest NOUN iajs-3467 275 16 . . PUNCT iajs-3467 276 1 r r NOUN iajs-3467 276 2 news news NOUN iajs-3467 276 3 . . PUNCT iajs-3467 277 1 2002 2002 NUM iajs-3467 277 2 , , PUNCT iajs-3467 277 3 3,18 3,18 NUM iajs-3467 277 4 - - SYM iajs-3467 277 5 22 22 NUM iajs-3467 277 6 . . PUNCT iajs-3467 278 1 3 3 X iajs-3467 278 2 . . X iajs-3467 278 3 khorshid khorshid PROPN iajs-3467 278 4 , , PUNCT iajs-3467 278 5 s.f s.f PROPN iajs-3467 278 6 . . PROPN iajs-3467 278 7 ; ; PUNCT iajs-3467 278 8 abdulazeez abdulazeez PROPN iajs-3467 278 9 , , PUNCT iajs-3467 278 10 a.m. a.m. PROPN iajs-3467 278 11 breast breast NOUN iajs-3467 278 12 cancer cancer NOUN iajs-3467 278 13 diagnosis diagnosis NOUN iajs-3467 278 14 based base VERB iajs-3467 278 15 on on ADP iajs-3467 278 16 k k ADJ iajs-3467 278 17 - - PUNCT iajs-3467 278 18 nearest near ADJ iajs-3467 278 19 neighbors neighbor NOUN iajs-3467 278 20 : : PUNCT iajs-3467 278 21 a a DET iajs-3467 278 22 review review NOUN iajs-3467 278 23 . . PUNCT iajs-3467 279 1 palarch palarch NOUN iajs-3467 279 2 's 's PART iajs-3467 279 3 journal journal NOUN iajs-3467 279 4 of of ADP iajs-3467 279 5 archaeology archaeology NOUN iajs-3467 279 6 of of ADP iajs-3467 279 7 egypt egypt PROPN iajs-3467 279 8 / / SYM iajs-3467 279 9 egyptology egyptology NOUN iajs-3467 279 10 . . PUNCT iajs-3467 280 1 2021 2021 NUM iajs-3467 280 2 , , PUNCT iajs-3467 280 3 18 18 NUM iajs-3467 280 4 , , PUNCT iajs-3467 280 5 1927 1927 NUM iajs-3467 280 6 - - SYM iajs-3467 280 7 51 51 NUM iajs-3467 280 8 . . PUNCT iajs-3467 281 1 4 4 NUM iajs-3467 281 2 . . X iajs-3467 282 1 chen chen PROPN iajs-3467 282 2 , , PUNCT iajs-3467 282 3 s. s. PROPN iajs-3467 282 4 ; ; PUNCT iajs-3467 282 5 webb webb PROPN iajs-3467 282 6 , , PUNCT iajs-3467 282 7 g.i g.i PROPN iajs-3467 282 8 . . PROPN iajs-3467 282 9 ; ; PUNCT iajs-3467 282 10 liu liu PROPN iajs-3467 282 11 , , PUNCT iajs-3467 282 12 l. l. PROPN iajs-3467 282 13 ; ; PUNCT iajs-3467 282 14 ma ma PROPN iajs-3467 282 15 , , PUNCT iajs-3467 282 16 x. x. NOUN iajs-3467 282 17 a. a. NOUN iajs-3467 282 18 novel novel PROPN iajs-3467 282 19 selective selective ADJ iajs-3467 282 20 naΓ―ve naΓ―ve ADJ iajs-3467 282 21 bayes bayes NOUN iajs-3467 282 22 algorithm algorithm PROPN iajs-3467 282 23 . . PUNCT iajs-3467 283 1 knowledgebased knowledgebased PROPN iajs-3467 283 2 systems system NOUN iajs-3467 283 3 . . PUNCT iajs-3467 284 1 2020 2020 NUM iajs-3467 284 2 , , PUNCT iajs-3467 284 3 192 192 NUM iajs-3467 284 4 , , PUNCT iajs-3467 284 5 105361 105361 NUM iajs-3467 284 6 5 5 NUM iajs-3467 284 7 . . PUNCT iajs-3467 285 1 choubey choubey PROPN iajs-3467 285 2 , , PUNCT iajs-3467 285 3 d.k d.k PROPN iajs-3467 285 4 . . PROPN iajs-3467 285 5 ; ; PUNCT iajs-3467 285 6 kumar kumar PROPN iajs-3467 285 7 , , PUNCT iajs-3467 285 8 m. m. NOUN iajs-3467 285 9 ; ; PUNCT iajs-3467 285 10 shukla shukla NOUN iajs-3467 285 11 , , PUNCT iajs-3467 285 12 v. v. PROPN iajs-3467 285 13 ; ; PUNCT iajs-3467 285 14 tripathi tripathi PROPN iajs-3467 285 15 , , PUNCT iajs-3467 285 16 s. s. PROPN iajs-3467 285 17 ; ; PUNCT iajs-3467 285 18 dhandhania dhandhania PROPN iajs-3467 285 19 , , PUNCT iajs-3467 285 20 v.k v.k PROPN iajs-3467 285 21 . . PROPN iajs-3467 285 22 ; ; PUNCT iajs-3467 285 23 comparative comparative ADJ iajs-3467 285 24 analysis analysis NOUN iajs-3467 285 25 of of ADP iajs-3467 285 26 classification classification NOUN iajs-3467 285 27 methods method NOUN iajs-3467 285 28 with with ADP iajs-3467 285 29 pca pca PROPN iajs-3467 285 30 and and CCONJ iajs-3467 285 31 lda lda VERB iajs-3467 285 32 for for ADP iajs-3467 285 33 diabetes diabetes NOUN iajs-3467 285 34 . . PUNCT iajs-3467 286 1 current current ADJ iajs-3467 286 2 diabetes diabetes NOUN iajs-3467 286 3 reviews review NOUN iajs-3467 286 4 . . PUNCT iajs-3467 287 1 2020 2020 NUM iajs-3467 287 2 , , PUNCT iajs-3467 287 3 16 16 NUM iajs-3467 287 4 , , PUNCT iajs-3467 287 5 833 833 NUM iajs-3467 287 6 - - SYM iajs-3467 287 7 50 50 NUM iajs-3467 287 8 . . NOUN iajs-3467 288 1 6 6 NUM iajs-3467 288 2 . . X iajs-3467 288 3 tawfiq tawfiq PROPN iajs-3467 288 4 , , PUNCT iajs-3467 288 5 l.n l.n PROPN iajs-3467 288 6 . . PROPN iajs-3467 288 7 ; ; PUNCT iajs-3467 288 8 rashid rashid PROPN iajs-3467 288 9 , , PUNCT iajs-3467 288 10 t.a t.a PROPN iajs-3467 288 11 . . PROPN iajs-3467 288 12 on on ADP iajs-3467 288 13 comparison comparison NOUN iajs-3467 288 14 between between ADP iajs-3467 288 15 radial radial ADJ iajs-3467 288 16 basis basis NOUN iajs-3467 288 17 function function NOUN iajs-3467 288 18 and and CCONJ iajs-3467 288 19 wavelet wavelet NOUN iajs-3467 288 20 basis basis NOUN iajs-3467 288 21 functions function NOUN iajs-3467 288 22 neural neural ADJ iajs-3467 288 23 networks network NOUN iajs-3467 288 24 . . PUNCT iajs-3467 289 1 ibn ibn PROPN iajs-3467 289 2 al al PROPN iajs-3467 289 3 - - PUNCT iajs-3467 289 4 haitham haitham PROPN iajs-3467 289 5 journal journal PROPN iajs-3467 289 6 for for ADP iajs-3467 289 7 pure pure ADJ iajs-3467 289 8 and and CCONJ iajs-3467 289 9 applied applied ADJ iajs-3467 289 10 science science NOUN iajs-3467 289 11 . . PUNCT iajs-3467 290 1 2017 2017 NUM iajs-3467 290 2 , , PUNCT iajs-3467 290 3 23 23 NUM iajs-3467 290 4 , , PUNCT iajs-3467 290 5 184 184 NUM iajs-3467 290 6 - - SYM iajs-3467 290 7 92 92 NUM iajs-3467 290 8 . . PUNCT iajs-3467 291 1 7 7 X iajs-3467 291 2 . . X iajs-3467 291 3 zhi zhi PROPN iajs-3467 291 4 , , PUNCT iajs-3467 291 5 j. j. PROPN iajs-3467 291 6 ; ; PUNCT iajs-3467 291 7 sun sun PROPN iajs-3467 291 8 , , PUNCT iajs-3467 291 9 j. j. PROPN iajs-3467 291 10 ; ; PUNCT iajs-3467 291 11 wang wang PROPN iajs-3467 291 12 , , PUNCT iajs-3467 291 13 z. z. PROPN iajs-3467 291 14 ; ; PUNCT iajs-3467 291 15 ding ding NOUN iajs-3467 291 16 , , PUNCT iajs-3467 291 17 w. w. PROPN iajs-3467 291 18 support support PROPN iajs-3467 291 19 vector vector NOUN iajs-3467 291 20 machine machine NOUN iajs-3467 291 21 classifier classifier NOUN iajs-3467 291 22 for for ADP iajs-3467 291 23 prediction prediction NOUN iajs-3467 291 24 of of ADP iajs-3467 291 25 the the DET iajs-3467 291 26 metastasis metastasis NOUN iajs-3467 291 27 of of ADP iajs-3467 291 28 colorectal colorectal ADJ iajs-3467 291 29 cancer cancer NOUN iajs-3467 291 30 . . PUNCT iajs-3467 292 1 int int PROPN iajs-3467 292 2 j j PROPN iajs-3467 292 3 mol mol ADP iajs-3467 292 4 med med PROPN iajs-3467 292 5 . . PROPN iajs-3467 292 6 2018 2018 NUM iajs-3467 292 7 , , PUNCT iajs-3467 292 8 41 41 NUM iajs-3467 292 9 , , PUNCT iajs-3467 292 10 1419 1419 NUM iajs-3467 292 11 - - SYM iajs-3467 292 12 26 26 NUM iajs-3467 292 13 . . NOUN iajs-3467 293 1 8 8 NUM iajs-3467 293 2 . . PUNCT iajs-3467 293 3 cervantes cervante NOUN iajs-3467 293 4 , , PUNCT iajs-3467 293 5 j. j. PROPN iajs-3467 293 6 ; ; PUNCT iajs-3467 293 7 garcia garcia PROPN iajs-3467 293 8 - - PUNCT iajs-3467 293 9 lamont lamont PROPN iajs-3467 293 10 , , PUNCT iajs-3467 293 11 f. f. PROPN iajs-3467 293 12 ; ; PUNCT iajs-3467 293 13 rodrΓ­guez rodrΓ­guez PROPN iajs-3467 293 14 - - PUNCT iajs-3467 293 15 mazahua mazahua PROPN iajs-3467 293 16 , , PUNCT iajs-3467 293 17 l. l. PROPN iajs-3467 293 18 ; ; PUNCT iajs-3467 293 19 lopez lopez NOUN iajs-3467 293 20 , , PUNCT iajs-3467 293 21 a. a. NOUN iajs-3467 293 22 a a DET iajs-3467 293 23 comprehensive comprehensive ADJ iajs-3467 293 24 survey survey NOUN iajs-3467 293 25 on on ADP iajs-3467 293 26 support support NOUN iajs-3467 293 27 vector vector NOUN iajs-3467 293 28 machine machine NOUN iajs-3467 293 29 classification classification NOUN iajs-3467 293 30 : : PUNCT iajs-3467 293 31 applications application NOUN iajs-3467 293 32 , , PUNCT iajs-3467 293 33 challenges challenge NOUN iajs-3467 293 34 and and CCONJ iajs-3467 293 35 trends trend NOUN iajs-3467 293 36 . . PUNCT iajs-3467 294 1 neurocomputing neurocompute VERB iajs-3467 294 2 . . PUNCT iajs-3467 295 1 2020 2020 NUM iajs-3467 295 2 , , PUNCT iajs-3467 295 3 408 408 NUM iajs-3467 295 4 , , PUNCT iajs-3467 295 5 189 189 NUM iajs-3467 295 6 - - SYM iajs-3467 295 7 215 215 NUM iajs-3467 295 8 . . NOUN iajs-3467 296 1 9 9 X iajs-3467 296 2 . . X iajs-3467 296 3 hekmatmanesh hekmatmanesh NOUN iajs-3467 296 4 , , PUNCT iajs-3467 296 5 a. a. NOUN iajs-3467 296 6 ; ; PUNCT iajs-3467 296 7 wu wu PROPN iajs-3467 296 8 , , PUNCT iajs-3467 296 9 h. h. PROPN iajs-3467 296 10 ; ; PUNCT iajs-3467 296 11 jamaloo jamaloo PROPN iajs-3467 296 12 , , PUNCT iajs-3467 296 13 f. f. PROPN iajs-3467 296 14 ; ; PUNCT iajs-3467 296 15 li li PROPN iajs-3467 296 16 , , PUNCT iajs-3467 296 17 m. m. NOUN iajs-3467 296 18 ; ; PUNCT iajs-3467 296 19 handroos handroos PROPN iajs-3467 296 20 , , PUNCT iajs-3467 296 21 h. h. PROPN iajs-3467 296 22 a a DET iajs-3467 296 23 combination combination NOUN iajs-3467 296 24 of of ADP iajs-3467 296 25 csp csp PROPN iajs-3467 296 26 - - PUNCT iajs-3467 296 27 based base VERB iajs-3467 296 28 method method NOUN iajs-3467 296 29 with with ADP iajs-3467 296 30 soft soft ADJ iajs-3467 296 31 margin margin NOUN iajs-3467 296 32 svm svm NOUN iajs-3467 296 33 classifier classifier NOUN iajs-3467 296 34 and and CCONJ iajs-3467 296 35 generalized generalize VERB iajs-3467 296 36 rbf rbf PROPN iajs-3467 296 37 kernel kernel PROPN iajs-3467 296 38 for for ADP iajs-3467 296 39 imagery imagery NOUN iajs-3467 296 40 - - PUNCT iajs-3467 296 41 based base VERB iajs-3467 296 42 brain brain NOUN iajs-3467 296 43 computer computer NOUN iajs-3467 296 44 interface interface NOUN iajs-3467 296 45 applications application NOUN iajs-3467 296 46 . . PUNCT iajs-3467 297 1 multimedia multimedia NOUN iajs-3467 297 2 tools tool NOUN iajs-3467 297 3 and and CCONJ iajs-3467 297 4 applications application NOUN iajs-3467 297 5 . . PUNCT iajs-3467 298 1 2020 2020 NUM iajs-3467 298 2 , , PUNCT iajs-3467 298 3 79 79 NUM iajs-3467 298 4 , , PUNCT iajs-3467 298 5 17521 17521 NUM iajs-3467 298 6 - - SYM iajs-3467 298 7 49 49 NUM iajs-3467 298 8 . . PUNCT iajs-3467 299 1 10 10 NUM iajs-3467 299 2 . . X iajs-3467 300 1 wang wang PROPN iajs-3467 300 2 , , PUNCT iajs-3467 300 3 y. y. PROPN iajs-3467 300 4 ; ; PUNCT iajs-3467 300 5 yu yu PROPN iajs-3467 300 6 , , PUNCT iajs-3467 300 7 w. w. PROPN iajs-3467 300 8 ; ; PUNCT iajs-3467 300 9 fang fang X iajs-3467 300 10 , , PUNCT iajs-3467 300 11 z. z. PROPN iajs-3467 300 12 multiple multiple ADJ iajs-3467 300 13 kernel kernel PROPN iajs-3467 300 14 - - PUNCT iajs-3467 300 15 based base VERB iajs-3467 300 16 svm svm ADJ iajs-3467 300 17 classification classification NOUN iajs-3467 300 18 of of ADP iajs-3467 300 19 hyperspectral hyperspectral ADJ iajs-3467 300 20 images image NOUN iajs-3467 300 21 by by ADP iajs-3467 300 22 combining combine VERB iajs-3467 300 23 spectral spectral ADJ iajs-3467 300 24 , , PUNCT iajs-3467 300 25 spatial spatial ADJ iajs-3467 300 26 , , PUNCT iajs-3467 300 27 and and CCONJ iajs-3467 300 28 semantic semantic ADJ iajs-3467 300 29 information information NOUN iajs-3467 300 30 . . PUNCT iajs-3467 301 1 remote remote ADJ iajs-3467 301 2 sensing sensing NOUN iajs-3467 301 3 . . PUNCT iajs-3467 302 1 2020 2020 NUM iajs-3467 302 2 , , PUNCT iajs-3467 302 3 12 12 NUM iajs-3467 302 4 , , PUNCT iajs-3467 302 5 120 120 NUM iajs-3467 302 6 . . PUNCT iajs-3467 302 7 ihjpas ihjpas PROPN iajs-3467 302 8 . . PUNCT iajs-3467 303 1 37 37 NUM iajs-3467 303 2 ( ( PUNCT iajs-3467 303 3 1 1 NUM iajs-3467 303 4 ) ) PUNCT iajs-3467 303 5 2024 2024 NUM iajs-3467 303 6 427 427 NUM iajs-3467 303 7 11 11 NUM iajs-3467 303 8 . . PUNCT iajs-3467 304 1 raheem raheem PROPN iajs-3467 304 2 , , PUNCT iajs-3467 304 3 s. s. PROPN iajs-3467 304 4 h h PROPN iajs-3467 304 5 ; ; PUNCT iajs-3467 304 6 kalaf kalaf PROPN iajs-3467 304 7 , , PUNCT iajs-3467 304 8 b. b. PROPN iajs-3467 304 9 a. a. PROPN iajs-3467 304 10 ; ; PUNCT iajs-3467 304 11 salman salman PROPN iajs-3467 304 12 , , PUNCT iajs-3467 304 13 a. a. PROPN iajs-3467 304 14 n. n. PROPN iajs-3467 304 15 comparison comparison NOUN iajs-3467 304 16 of of ADP iajs-3467 304 17 some some PRON iajs-3467 304 18 of of ADP iajs-3467 304 19 estimation estimation NOUN iajs-3467 304 20 methods method NOUN iajs-3467 304 21 of of ADP iajs-3467 304 22 stress stress NOUN iajs-3467 304 23 - - PUNCT iajs-3467 304 24 strength strength NOUN iajs-3467 304 25 model model NOUN iajs-3467 304 26 : : PUNCT iajs-3467 304 27 r= r= PROPN iajs-3467 304 28 p p X iajs-3467 304 29 ( ( PUNCT iajs-3467 304 30 y y X iajs-3467 304 31 < < X iajs-3467 304 32 x x X iajs-3467 304 33 < < X iajs-3467 304 34 z z PROPN iajs-3467 304 35 ) ) PUNCT iajs-3467 304 36 . . PUNCT iajs-3467 305 1 baghdad baghdad PROPN iajs-3467 305 2 science science PROPN iajs-3467 305 3 journal journal PROPN iajs-3467 305 4 , , PUNCT iajs-3467 305 5 2021 2021 NUM iajs-3467 305 6 , , PUNCT iajs-3467 305 7 18.2 18.2 NUM iajs-3467 305 8 , , PUNCT iajs-3467 305 9 1103 1103 NUM iajs-3467 305 10 - - SYM iajs-3467 305 11 1103 1103 NUM iajs-3467 305 12 . . PUNCT iajs-3467 306 1 12 12 NUM iajs-3467 306 2 . . PUNCT iajs-3467 307 1 jebur jebur PROPN iajs-3467 307 2 , , PUNCT iajs-3467 307 3 i. i. PROPN iajs-3467 307 4 g. g. PROPN iajs-3467 307 5 ; ; PUNCT iajs-3467 307 6 kalaf kalaf PROPN iajs-3467 307 7 , , PUNCT iajs-3467 307 8 b. b. PROPN iajs-3467 307 9 a. a. PROPN iajs-3467 307 10 ; ; PUNCT iajs-3467 307 11 salman salman PROPN iajs-3467 307 12 , , PUNCT iajs-3467 307 13 a. a. PROPN iajs-3467 307 14 n. n. PROPN iajs-3467 307 15 an an DET iajs-3467 307 16 efficient efficient ADJ iajs-3467 307 17 shrinkage shrinkage NOUN iajs-3467 307 18 estimators estimator NOUN iajs-3467 307 19 for for ADP iajs-3467 307 20 generalized generalized ADJ iajs-3467 307 21 inverse inverse NOUN iajs-3467 307 22 rayleigh rayleigh NOUN iajs-3467 307 23 distribution distribution NOUN iajs-3467 307 24 based base VERB iajs-3467 307 25 on on ADP iajs-3467 307 26 bounded bounded ADJ iajs-3467 307 27 and and CCONJ iajs-3467 307 28 series series VERB iajs-3467 307 29 stress stress NOUN iajs-3467 307 30 - - PUNCT iajs-3467 307 31 strength strength NOUN iajs-3467 307 32 models model NOUN iajs-3467 307 33 . . PUNCT iajs-3467 308 1 in in ADP iajs-3467 308 2 : : PUNCT iajs-3467 308 3 journal journal PROPN iajs-3467 308 4 of of ADP iajs-3467 308 5 physics physics PROPN iajs-3467 308 6 : : PUNCT iajs-3467 308 7 conference conference NOUN iajs-3467 308 8 series series NOUN iajs-3467 308 9 . . PUNCT iajs-3467 309 1 iop iop PROPN iajs-3467 309 2 publishing publishing NOUN iajs-3467 309 3 , , PUNCT iajs-3467 309 4 2021 2021 NUM iajs-3467 309 5 , , PUNCT iajs-3467 309 6 012054 012054 NUM iajs-3467 309 7 . . PUNCT iajs-3467 310 1 13 13 NUM iajs-3467 310 2 . . X iajs-3467 311 1 mahdi mahdi PROPN iajs-3467 311 2 , , PUNCT iajs-3467 311 3 g.j g.j PROPN iajs-3467 311 4 . . PROPN iajs-3467 311 5 ; ; PUNCT iajs-3467 311 6 mohammed mohammed PROPN iajs-3467 311 7 , , PUNCT iajs-3467 311 8 n.j n.j PROPN iajs-3467 311 9 . . PROPN iajs-3467 311 10 ; ; PUNCT iajs-3467 312 1 al al PROPN iajs-3467 312 2 - - PUNCT iajs-3467 312 3 sharea sharea PROPN iajs-3467 312 4 , , PUNCT iajs-3467 312 5 z.i z.i PROPN iajs-3467 312 6 . . PROPN iajs-3467 312 7 regression regression NOUN iajs-3467 312 8 shrinkage shrinkage NOUN iajs-3467 312 9 and and CCONJ iajs-3467 312 10 selection selection NOUN iajs-3467 312 11 variables variable NOUN iajs-3467 312 12 via via ADP iajs-3467 312 13 an an DET iajs-3467 312 14 adaptive adaptive ADJ iajs-3467 312 15 elastic elastic ADJ iajs-3467 312 16 net net ADJ iajs-3467 312 17 model model NOUN iajs-3467 312 18 . . PUNCT iajs-3467 313 1 in in ADP iajs-3467 313 2 journal journal PROPN iajs-3467 313 3 of of ADP iajs-3467 313 4 physics physics PROPN iajs-3467 313 5 : : PUNCT iajs-3467 313 6 conference conference NOUN iajs-3467 313 7 series series NOUN iajs-3467 313 8 2021 2021 NUM iajs-3467 313 9 , , PUNCT iajs-3467 313 10 1879 1879 NUM iajs-3467 313 11 , , PUNCT iajs-3467 313 12 032014 032014 NUM iajs-3467 313 13 . . PUNCT iajs-3467 314 1 14 14 NUM iajs-3467 314 2 . . PUNCT iajs-3467 315 1 qingyang qingyang PROPN iajs-3467 315 2 , , PUNCT iajs-3467 315 3 z. z. PROPN iajs-3467 315 4 ; ; PUNCT iajs-3467 315 5 ghadeer ghadeer NOUN iajs-3467 315 6 , , PUNCT iajs-3467 315 7 m. m. NOUN iajs-3467 315 8 ; ; PUNCT iajs-3467 315 9 jian jian PROPN iajs-3467 315 10 , , PUNCT iajs-3467 315 11 t. t. PROPN iajs-3467 315 12 ; ; PUNCT iajs-3467 315 13 hao hao PROPN iajs-3467 315 14 , , PUNCT iajs-3467 315 15 c. c. PROPN iajs-3467 315 16 ; ; PUNCT iajs-3467 315 17 a a DET iajs-3467 315 18 graph graph NOUN iajs-3467 315 19 - - PUNCT iajs-3467 315 20 based base VERB iajs-3467 315 21 multi multi ADJ iajs-3467 315 22 - - ADJ iajs-3467 315 23 sample sample ADJ iajs-3467 315 24 test test NOUN iajs-3467 315 25 for for ADP iajs-3467 315 26 identifying identify VERB iajs-3467 315 27 pathways pathway NOUN iajs-3467 315 28 associated associate VERB iajs-3467 315 29 with with ADP iajs-3467 315 30 cancer cancer NOUN iajs-3467 315 31 progression progression NOUN iajs-3467 315 32 . . PUNCT iajs-3467 316 1 computational computational ADJ iajs-3467 316 2 biology biology NOUN iajs-3467 316 3 and and CCONJ iajs-3467 316 4 chemistry chemistry NOUN iajs-3467 316 5 , , PUNCT iajs-3467 316 6 2020 2020 NUM iajs-3467 316 7 , , PUNCT iajs-3467 316 8 87 87 NUM iajs-3467 316 9 : : SYM iajs-3467 316 10 107285 107285 NUM iajs-3467 316 11 . . PUNCT iajs-3467 317 1 15 15 NUM iajs-3467 317 2 . . X iajs-3467 318 1 zhang zhang PROPN iajs-3467 318 2 , , PUNCT iajs-3467 318 3 q. q. PROPN iajs-3467 318 4 ; ; PUNCT iajs-3467 318 5 dao dao PROPN iajs-3467 318 6 , , PUNCT iajs-3467 318 7 t. t. NOUN iajs-3467 318 8 a a DET iajs-3467 318 9 distance distance NOUN iajs-3467 318 10 based base VERB iajs-3467 318 11 multisampling multisampling NOUN iajs-3467 318 12 test test NOUN iajs-3467 318 13 for for ADP iajs-3467 318 14 high high ADJ iajs-3467 318 15 - - PUNCT iajs-3467 318 16 dimensional dimensional ADJ iajs-3467 318 17 compositional compositional ADJ iajs-3467 318 18 data datum NOUN iajs-3467 318 19 with with ADP iajs-3467 318 20 applications application NOUN iajs-3467 318 21 to to ADP iajs-3467 318 22 the the DET iajs-3467 318 23 human human ADJ iajs-3467 318 24 microbiome microbiome NOUN iajs-3467 318 25 . . PUNCT iajs-3467 319 1 bmc bmc ADJ iajs-3467 319 2 bioinformatics bioinformatics PROPN iajs-3467 319 3 , , PUNCT iajs-3467 319 4 2020 2020 NUM iajs-3467 319 5 , , PUNCT iajs-3467 319 6 21 21 NUM iajs-3467 319 7 , , PUNCT iajs-3467 319 8 1 1 NUM iajs-3467 319 9 - - SYM iajs-3467 319 10 17 17 NUM iajs-3467 319 11 . . PUNCT iajs-3467 320 1 16 16 NUM iajs-3467 320 2 . . PUNCT iajs-3467 321 1 mahdi mahdi PROPN iajs-3467 321 2 , , PUNCT iajs-3467 321 3 g.j g.j PROPN iajs-3467 321 4 , , PUNCT iajs-3467 321 5 kalaf kalaf PROPN iajs-3467 321 6 , , PUNCT iajs-3467 321 7 b.a b.a PROPN iajs-3467 321 8 . . PROPN iajs-3467 321 9 ; ; PUNCT iajs-3467 321 10 khaleel khaleel PROPN iajs-3467 321 11 , , PUNCT iajs-3467 321 12 m.a m.a PROPN iajs-3467 321 13 . . PROPN iajs-3467 321 14 enhanced enhance VERB iajs-3467 321 15 supervised supervised ADJ iajs-3467 321 16 principal principal ADJ iajs-3467 321 17 component component NOUN iajs-3467 321 18 analysis analysis NOUN iajs-3467 321 19 for for ADP iajs-3467 321 20 cancer cancer NOUN iajs-3467 321 21 classification classification NOUN iajs-3467 321 22 . . PUNCT iajs-3467 322 1 iraqi iraqi ADJ iajs-3467 322 2 journal journal PROPN iajs-3467 322 3 of of ADP iajs-3467 322 4 science science NOUN iajs-3467 322 5 . . PUNCT iajs-3467 323 1 2021 2021 NUM iajs-3467 323 2 , , PUNCT iajs-3467 323 3 1321 1321 NUM iajs-3467 323 4 - - SYM iajs-3467 323 5 33 33 NUM iajs-3467 323 6 . . PUNCT iajs-3467 324 1 17 17 NUM iajs-3467 324 2 . . PUNCT iajs-3467 325 1 mseer mseer NOUN iajs-3467 325 2 , , PUNCT iajs-3467 325 3 h.a h.a PROPN iajs-3467 325 4 . . PROPN iajs-3467 325 5 ; ; PUNCT iajs-3467 326 1 mahdi mahdi PROPN iajs-3467 326 2 , , PUNCT iajs-3467 326 3 g.j g.j PROPN iajs-3467 326 4 . . PROPN iajs-3467 326 5 comparison comparison NOUN iajs-3467 326 6 among among ADP iajs-3467 326 7 variable variable ADJ iajs-3467 326 8 selection selection NOUN iajs-3467 326 9 models model NOUN iajs-3467 326 10 and and CCONJ iajs-3467 326 11 its its PRON iajs-3467 326 12 application application NOUN iajs-3467 326 13 to to ADP iajs-3467 326 14 health health NOUN iajs-3467 326 15 dataset dataset NOUN iajs-3467 326 16 . . PUNCT iajs-3467 327 1 inaip inaip PROPN iajs-3467 327 2 conference conference NOUN iajs-3467 327 3 proceedings proceeding NOUN iajs-3467 327 4 2023 2023 NUM iajs-3467 327 5 , , PUNCT iajs-3467 327 6 1 1 NUM iajs-3467 327 7 , , PUNCT iajs-3467 327 8 2414 2414 NUM iajs-3467 327 9 . . PUNCT iajs-3467 328 1 18 18 NUM iajs-3467 328 2 . . PUNCT iajs-3467 329 1 jabbar jabbar PROPN iajs-3467 329 2 , , PUNCT iajs-3467 329 3 a.k a.k PROPN iajs-3467 329 4 . . PROPN iajs-3467 329 5 new new ADJ iajs-3467 329 6 transform transform VERB iajs-3467 329 7 fundamental fundamental ADJ iajs-3467 329 8 properties property NOUN iajs-3467 329 9 and and CCONJ iajs-3467 329 10 its its PRON iajs-3467 329 11 applications application NOUN iajs-3467 329 12 . . PUNCT iajs-3467 330 1 ibn ibn PROPN iajs-3467 330 2 al al PROPN iajs-3467 330 3 - - PUNCT iajs-3467 330 4 haitham haitham PROPN iajs-3467 330 5 journal journal PROPN iajs-3467 330 6 for for ADP iajs-3467 330 7 pure pure ADJ iajs-3467 330 8 and and CCONJ iajs-3467 330 9 applied applied ADJ iajs-3467 330 10 sciences science NOUN iajs-3467 330 11 . . PUNCT iajs-3467 330 12 2018 2018 NUM iajs-3467 330 13 , , PUNCT iajs-3467 330 14 31 31 NUM iajs-3467 330 15 , , PUNCT iajs-3467 330 16 1 1 NUM iajs-3467 330 17 - - SYM iajs-3467 330 18 10 10 NUM iajs-3467 330 19 . . PUNCT iajs-3467 330 20 19 19 NUM iajs-3467 330 21 . . X iajs-3467 331 1 mahdi mahdi PROPN iajs-3467 331 2 , , PUNCT iajs-3467 331 3 g.j g.j PROPN iajs-3467 331 4 . . PROPN iajs-3467 331 5 ; ; PUNCT iajs-3467 331 6 a a DET iajs-3467 331 7 modified modify VERB iajs-3467 331 8 support support NOUN iajs-3467 331 9 vector vector NOUN iajs-3467 331 10 machine machine NOUN iajs-3467 331 11 classifiers classifier NOUN iajs-3467 331 12 using use VERB iajs-3467 331 13 stochastic stochastic ADJ iajs-3467 331 14 gradient gradient ADJ iajs-3467 331 15 descent descent NOUN iajs-3467 331 16 with with ADP iajs-3467 331 17 application application NOUN iajs-3467 331 18 to to ADP iajs-3467 331 19 leukemia leukemia NOUN iajs-3467 331 20 cancer cancer NOUN iajs-3467 331 21 type type NOUN iajs-3467 331 22 dataset dataset NOUN iajs-3467 331 23 . . PUNCT iajs-3467 332 1 baghdad baghdad PROPN iajs-3467 332 2 science science PROPN iajs-3467 332 3 journal journal PROPN iajs-3467 332 4 . . PUNCT iajs-3467 333 1 2020,17,1255 2020,17,1255 NUM iajs-3467 333 2 - - SYM iajs-3467 333 3 69 69 NUM iajs-3467 333 4 . . PUNCT iajs-3467 334 1 20 20 NUM iajs-3467 334 2 . . PUNCT iajs-3467 335 1 raheem raheem PROPN iajs-3467 335 2 , , PUNCT iajs-3467 335 3 s.h s.h PROPN iajs-3467 335 4 . . PROPN iajs-3467 335 5 ; ; PUNCT iajs-3467 335 6 kalaf kalaf PROPN iajs-3467 335 7 , , PUNCT iajs-3467 335 8 b.a b.a PROPN iajs-3467 335 9 . . PROPN iajs-3467 335 10 ; ; PUNCT iajs-3467 335 11 salman salman PROPN iajs-3467 335 12 a.n a.n PROPN iajs-3467 335 13 . . PROPN iajs-3467 335 14 comparison comparison NOUN iajs-3467 335 15 of of ADP iajs-3467 335 16 some some PRON iajs-3467 335 17 of of ADP iajs-3467 335 18 estimation estimation NOUN iajs-3467 335 19 methods method NOUN iajs-3467 335 20 of of ADP iajs-3467 335 21 stress stress NOUN iajs-3467 335 22 - - PUNCT iajs-3467 335 23 strength strength NOUN iajs-3467 335 24 model model NOUN iajs-3467 335 25 : : PUNCT iajs-3467 335 26 r= r= PROPN iajs-3467 335 27 p p X iajs-3467 335 28 ( ( PUNCT iajs-3467 335 29 y y X iajs-3467 335 30 < < X iajs-3467 335 31 x x X iajs-3467 335 32 < < X iajs-3467 335 33 z z PROPN iajs-3467 335 34 ) ) PUNCT iajs-3467 335 35 . . PUNCT iajs-3467 336 1 baghdad baghdad PROPN iajs-3467 336 2 science science PROPN iajs-3467 336 3 journal journal PROPN iajs-3467 336 4 . . PUNCT iajs-3467 337 1 2021,18,1103 2021,18,1103 NUM iajs-3467 337 2 - - SYM iajs-3467 337 3 17 17 NUM iajs-3467 337 4 . . PUNCT iajs-3467 338 1 21 21 NUM iajs-3467 338 2 . . PUNCT iajs-3467 339 1 salah salah PROPN iajs-3467 339 2 , , PUNCT iajs-3467 339 3 o.m o.m PROPN iajs-3467 339 4 . . PROPN iajs-3467 339 5 ; ; PUNCT iajs-3467 339 6 mahdi mahdi PROPN iajs-3467 339 7 , , PUNCT iajs-3467 339 8 g.j g.j PROPN iajs-3467 339 9 . . PROPN iajs-3467 339 10 ; ; PUNCT iajs-3467 340 1 al al PROPN iajs-3467 340 2 - - PUNCT iajs-3467 340 3 latif latif PROPN iajs-3467 340 4 , , PUNCT iajs-3467 340 5 i.a i.a PROPN iajs-3467 340 6 . . PROPN iajs-3467 340 7 a a DET iajs-3467 340 8 modified modify VERB iajs-3467 340 9 arima arima NOUN iajs-3467 340 10 model model NOUN iajs-3467 340 11 for for ADP iajs-3467 340 12 forecasting forecast VERB iajs-3467 340 13 chemical chemical ADJ iajs-3467 340 14 sales sale NOUN iajs-3467 340 15 in in ADP iajs-3467 340 16 the the DET iajs-3467 340 17 usa usa PROPN iajs-3467 340 18 . . PROPN iajs-3467 340 19 in in ADP iajs-3467 340 20 journal journal PROPN iajs-3467 340 21 of of ADP iajs-3467 340 22 physics physics PROPN iajs-3467 340 23 : : PUNCT iajs-3467 340 24 conference conference NOUN iajs-3467 340 25 series series NOUN iajs-3467 340 26 2021 2021 NUM iajs-3467 340 27 , , PUNCT iajs-3467 340 28 1879 1879 NUM iajs-3467 340 29 , , PUNCT iajs-3467 340 30 032008 032008 NUM iajs-3467 340 31 . . PUNCT iajs-3467 341 1 22 22 NUM iajs-3467 341 2 . . PUNCT iajs-3467 342 1 al al PROPN iajs-3467 342 2 - - PUNCT iajs-3467 342 3 noor noor PROPN iajs-3467 342 4 , , PUNCT iajs-3467 342 5 n. n. PROPN iajs-3467 342 6 h. h. PROPN iajs-3467 342 7 ; ; PUNCT iajs-3467 342 8 khaleel khaleel PROPN iajs-3467 342 9 , , PUNCT iajs-3467 342 10 m. m. NOUN iajs-3467 342 11 a. a. PROPN iajs-3467 342 12 ; ; PUNCT iajs-3467 342 13 mohammed mohammed PROPN iajs-3467 342 14 , , PUNCT iajs-3467 342 15 g. g. PROPN iajs-3467 342 16 j. j. PROPN iajs-3467 342 17 theory theory PROPN iajs-3467 342 18 and and CCONJ iajs-3467 342 19 applications application NOUN iajs-3467 342 20 of of ADP iajs-3467 342 21 marshall marshall PROPN iajs-3467 342 22 olkin olkin PROPN iajs-3467 342 23 marshall marshall PROPN iajs-3467 342 24 olkin olkin PROPN iajs-3467 342 25 weibull weibull PROPN iajs-3467 342 26 distribution distribution NOUN iajs-3467 342 27 . . PUNCT iajs-3467 343 1 in in ADP iajs-3467 343 2 : : PUNCT iajs-3467 343 3 journal journal PROPN iajs-3467 343 4 of of ADP iajs-3467 343 5 physics physics PROPN iajs-3467 343 6 : : PUNCT iajs-3467 343 7 conference conference NOUN iajs-3467 343 8 series series NOUN iajs-3467 343 9 . . PUNCT iajs-3467 344 1 2021,20 2021,20 PROPN iajs-3467 344 2 , , PUNCT iajs-3467 344 3 012101 012101 NUM iajs-3467 344 4 . . PUNCT iajs-3467 345 1 23 23 NUM iajs-3467 345 2 . . PUNCT iajs-3467 346 1 sheah sheah PROPN iajs-3467 346 2 , , PUNCT iajs-3467 346 3 r. r. PROPN iajs-3467 346 4 h. h. PROPN iajs-3467 346 5 ; ; PUNCT iajs-3467 346 6 abbas abbas PROPN iajs-3467 346 7 , , PUNCT iajs-3467 346 8 i. i. PROPN iajs-3467 346 9 t. t. PROPN iajs-3467 346 10 using use VERB iajs-3467 346 11 multi multi ADJ iajs-3467 346 12 - - ADJ iajs-3467 346 13 objective objective ADJ iajs-3467 346 14 bat bat NOUN iajs-3467 346 15 algorithm algorithm NOUN iajs-3467 346 16 for for ADP iajs-3467 346 17 solving solve VERB iajs-3467 346 18 multi multi ADJ iajs-3467 346 19 - - ADJ iajs-3467 346 20 objective objective ADJ iajs-3467 346 21 non non ADJ iajs-3467 346 22 - - ADJ iajs-3467 346 23 linear linear ADJ iajs-3467 346 24 programming programming NOUN iajs-3467 346 25 problem problem NOUN iajs-3467 346 26 . . PUNCT iajs-3467 347 1 iraqi iraqi ADJ iajs-3467 347 2 journal journal PROPN iajs-3467 347 3 of of ADP iajs-3467 347 4 science science NOUN iajs-3467 347 5 , , PUNCT iajs-3467 347 6 2021 2021 NUM iajs-3467 347 7 , , PUNCT iajs-3467 347 8 997 997 NUM iajs-3467 347 9 - - SYM iajs-3467 347 10 1015 1015 NUM iajs-3467 347 11 . . PUNCT iajs-3467 348 1 24 24 NUM iajs-3467 348 2 . . PUNCT iajs-3467 349 1 mohammed mohammed PROPN iajs-3467 349 2 , , PUNCT iajs-3467 349 3 m. m. NOUN iajs-3467 349 4 j. j. PROPN iajs-3467 349 5 ; ; PUNCT iajs-3467 349 6 mohammed mohammed PROPN iajs-3467 349 7 , , PUNCT iajs-3467 349 8 a. a. NOUN iajs-3467 349 9 t. t. NOUN iajs-3467 349 10 analysis analysis NOUN iajs-3467 349 11 of of ADP iajs-3467 349 12 an an DET iajs-3467 349 13 agriculture agriculture NOUN iajs-3467 349 14 data datum NOUN iajs-3467 349 15 using use VERB iajs-3467 349 16 markov markov NOUN iajs-3467 349 17 basis basis NOUN iajs-3467 349 18 for for ADP iajs-3467 349 19 independent independent ADJ iajs-3467 349 20 model model NOUN iajs-3467 349 21 . . PUNCT iajs-3467 350 1 in in ADP iajs-3467 350 2 : : PUNCT iajs-3467 350 3 journal journal PROPN iajs-3467 350 4 of of ADP iajs-3467 350 5 physics physics PROPN iajs-3467 350 6 : : PUNCT iajs-3467 350 7 conference conference NOUN iajs-3467 350 8 series series NOUN iajs-3467 350 9 . . PUNCT iajs-3467 351 1 2020 2020 NUM iajs-3467 351 2 , , PUNCT iajs-3467 351 3 012071 012071 NUM iajs-3467 351 4 . . PUNCT iajs-3467 352 1 25 25 NUM iajs-3467 352 2 . . PUNCT iajs-3467 353 1 mohammed mohammed PROPN iajs-3467 353 2 , , PUNCT iajs-3467 353 3 m. m. NOUN iajs-3467 353 4 j. j. PROPN iajs-3467 353 5 ; ; PUNCT iajs-3467 353 6 mohammed mohammed PROPN iajs-3467 353 7 , , PUNCT iajs-3467 353 8 a. a. NOUN iajs-3467 353 9 t. t. PROPN iajs-3467 353 10 parameter parameter PROPN iajs-3467 353 11 estimation estimation NOUN iajs-3467 353 12 of of ADP iajs-3467 353 13 inverse inverse NOUN iajs-3467 353 14 exponential exponential ADJ iajs-3467 353 15 rayleigh rayleigh NOUN iajs-3467 353 16 distribution distribution NOUN iajs-3467 353 17 based base VERB iajs-3467 353 18 on on ADP iajs-3467 353 19 classical classical ADJ iajs-3467 353 20 methods method NOUN iajs-3467 353 21 . . PUNCT iajs-3467 354 1 international international ADJ iajs-3467 354 2 journal journal PROPN iajs-3467 354 3 of of ADP iajs-3467 354 4 nonlinear nonlinear ADJ iajs-3467 354 5 analysis analysis NOUN iajs-3467 354 6 and and CCONJ iajs-3467 354 7 applications application NOUN iajs-3467 354 8 , , PUNCT iajs-3467 354 9 2021 2021 NUM iajs-3467 354 10 , , PUNCT iajs-3467 354 11 12 12 NUM iajs-3467 354 12 , , PUNCT iajs-3467 354 13 935 935 NUM iajs-3467 354 14 - - SYM iajs-3467 354 15 944 944 NUM iajs-3467 354 16 . . NOUN iajs-3467 354 17 26 26 NUM iajs-3467 354 18 . . PUNCT iajs-3467 355 1 bartley bartley PROPN iajs-3467 355 2 , , PUNCT iajs-3467 355 3 c. c. PROPN iajs-3467 355 4 replication replication NOUN iajs-3467 355 5 data datum NOUN iajs-3467 355 6 for for ADP iajs-3467 355 7 : : PUNCT iajs-3467 355 8 south south ADJ iajs-3467 355 9 african african ADJ iajs-3467 355 10 heart heart NOUN iajs-3467 355 11 disease disease NOUN iajs-3467 355 12 " " PUNCT iajs-3467 355 13 available available ADJ iajs-3467 355 14 online online ADV iajs-3467 355 15 : : PUNCT iajs-3467 355 16 https://doi.org/10.7910/dvn/76siqd https://doi.org/10.7910/dvn/76siqd PROPN iajs-3467 355 17 harvard harvard PROPN iajs-3467 355 18 dataverse dataverse PROPN iajs-3467 355 19 , , PUNCT iajs-3467 355 20 v1 v1 NOUN iajs-3467 355 21 , , PUNCT iajs-3467 355 22 2016 2016 NUM iajs-3467 355 23 . . PUNCT iajs-3467 356 1 27 27 NUM iajs-3467 356 2 . . PUNCT iajs-3467 357 1 bayda bayda VERB iajs-3467 357 2 , , PUNCT iajs-3467 357 3 a. a. PROPN iajs-3467 357 4 abdul abdul PROPN iajs-3467 357 5 jabbar jabbar PROPN iajs-3467 357 6 , , PUNCT iajs-3467 357 7 k. k. PROPN iajs-3467 357 8 b. b. PROPN iajs-3467 357 9 ; ; PUNCT iajs-3467 357 10 iraq iraq PROPN iajs-3467 357 11 , , PUNCT iajs-3467 357 12 t. t. NOUN iajs-3467 357 13 a. a. NOUN iajs-3467 357 14 mohd mohd PROPN iajs-3467 357 15 , , PUNCT iajs-3467 357 16 r. r. PROPN iajs-3467 357 17 a. a. PROPN iajs-3467 357 18 ; ; PUNCT iajs-3467 357 19 lee lee PROPN iajs-3467 357 20 , , PUNCT iajs-3467 357 21 l. l. PROPN iajs-3467 357 22 s. s. PROPN iajs-3467 357 23 application application PROPN iajs-3467 357 24 of of ADP iajs-3467 357 25 simulated simulate VERB iajs-3467 357 26 annealing annealing NOUN iajs-3467 357 27 to to PART iajs-3467 357 28 solve solve VERB iajs-3467 357 29 multi multi NOUN iajs-3467 357 30 - - NOUN iajs-3467 357 31 objectives objective NOUN iajs-3467 357 32 for for ADP iajs-3467 357 33 aggregate aggregate ADJ iajs-3467 357 34 production production NOUN iajs-3467 357 35 planning planning NOUN iajs-3467 357 36 . . PUNCT iajs-3467 358 1 in in ADP iajs-3467 358 2 : : PUNCT iajs-3467 358 3 aip aip PROPN iajs-3467 358 4 conference conference NOUN iajs-3467 358 5 proceedings proceeding NOUN iajs-3467 358 6 . . PUNCT iajs-3467 359 1 2016 2016 NUM iajs-3467 359 2 , , PUNCT iajs-3467 359 3 1739 1739 NUM iajs-3467 359 4 , , PUNCT iajs-3467 359 5 020086 020086 NUM iajs-3467 359 6 . . PUNCT iajs-3467 360 1 28 28 NUM iajs-3467 360 2 . . PUNCT iajs-3467 360 3 bogatinovski bogatinovski PROPN iajs-3467 360 4 , , PUNCT iajs-3467 360 5 j. j. PROPN iajs-3467 360 6 ; ; PUNCT iajs-3467 360 7 ljupčo ljupčo PROPN iajs-3467 360 8 , , PUNCT iajs-3467 360 9 t. t. PROPN iajs-3467 360 10 ; ; PUNCT iajs-3467 360 11 saΕ‘o saΕ‘o PROPN iajs-3467 360 12 , , PUNCT iajs-3467 360 13 d. d. PROPN iajs-3467 360 14 ; ; PUNCT iajs-3467 360 15 dragi dragi PROPN iajs-3467 360 16 , , PUNCT iajs-3467 360 17 kocev kocev PROPN iajs-3467 360 18 . . PUNCT iajs-3467 361 1 comprehensive comprehensive ADJ iajs-3467 361 2 comparative comparative ADJ iajs-3467 361 3 study study NOUN iajs-3467 361 4 of of ADP iajs-3467 361 5 multi multi ADJ iajs-3467 361 6 - - ADJ iajs-3467 361 7 label label ADJ iajs-3467 361 8 classification classification NOUN iajs-3467 361 9 methods method NOUN iajs-3467 361 10 . . PUNCT iajs-3467 362 1 expert expert NOUN iajs-3467 362 2 systems system NOUN iajs-3467 362 3 with with ADP iajs-3467 362 4 applications application NOUN iajs-3467 362 5 . . PUNCT iajs-3467 363 1 2022 2022 NUM iajs-3467 363 2 , , PUNCT iajs-3467 363 3 203 203 NUM iajs-3467 363 4 , , PUNCT iajs-3467 363 5 117215 117215 NUM iajs-3467 363 6 . . PUNCT iajs-3467 364 1 https://doi.org/10.7910/dvn/76siqd https://doi.org/10.7910/dvn/76siqd NOUN iajs-3467 364 2 https://doi.org/10.7910/dvn/76siqd https://doi.org/10.7910/dvn/76siqd NOUN iajs-3467 364 3 ihjpas ihjpa NOUN iajs-3467 364 4 . . PUNCT iajs-3467 365 1 37 37 NUM iajs-3467 365 2 ( ( PUNCT iajs-3467 365 3 1 1 NUM iajs-3467 365 4 ) ) PUNCT iajs-3467 365 5 2024 2024 NUM iajs-3467 365 6 428 428 NUM iajs-3467 365 7 29 29 NUM iajs-3467 365 8 . . PUNCT iajs-3467 366 1 fjellstrΓΆm fjellstrΓΆm NOUN iajs-3467 366 2 , , PUNCT iajs-3467 366 3 c. c. PROPN iajs-3467 366 4 ; ; PUNCT iajs-3467 366 5 nystrΓΆm nystrΓΆm PROPN iajs-3467 366 6 , , PUNCT iajs-3467 366 7 kaj kaj PROPN iajs-3467 366 8 . . PUNCT iajs-3467 367 1 deep deep ADJ iajs-3467 367 2 learning learning NOUN iajs-3467 367 3 , , PUNCT iajs-3467 367 4 stochastic stochastic ADJ iajs-3467 367 5 gradient gradient ADJ iajs-3467 367 6 descent descent NOUN iajs-3467 367 7 and and CCONJ iajs-3467 367 8 diffusion diffusion NOUN iajs-3467 367 9 maps map NOUN iajs-3467 367 10 . . PUNCT iajs-3467 368 1 journal journal NOUN iajs-3467 368 2 of of ADP iajs-3467 368 3 computational computational ADJ iajs-3467 368 4 mathematics mathematic NOUN iajs-3467 368 5 and and CCONJ iajs-3467 368 6 data datum NOUN iajs-3467 368 7 science science NOUN iajs-3467 368 8 . . PUNCT iajs-3467 369 1 2022 2022 NUM iajs-3467 369 2 , , PUNCT iajs-3467 369 3 4 4 NUM iajs-3467 369 4 , , PUNCT iajs-3467 369 5 100054 100054 NUM iajs-3467 369 6 . . PUNCT iajs-3467 370 1 30 30 NUM iajs-3467 370 2 . . PUNCT iajs-3467 371 1 hassan hassan PROPN iajs-3467 371 2 , , PUNCT iajs-3467 371 3 a. a. PROPN iajs-3467 371 4 s. s. PROPN iajs-3467 371 5 ; ; PUNCT iajs-3467 371 6 khaleel khaleel PROPN iajs-3467 371 7 , , PUNCT iajs-3467 371 8 m. m. NOUN iajs-3467 371 9 a. a. NOUN iajs-3467 371 10 ; ; PUNCT iajs-3467 371 11 mohamd mohamd PROPN iajs-3467 371 12 , , PUNCT iajs-3467 371 13 r. r. PROPN iajs-3467 371 14 e. e. PROPN iajs-3467 372 1 an an DET iajs-3467 372 2 extension extension NOUN iajs-3467 372 3 of of ADP iajs-3467 372 4 exponentiated exponentiated ADJ iajs-3467 372 5 lomax lomax PROPN iajs-3467 372 6 distribution distribution NOUN iajs-3467 372 7 with with ADP iajs-3467 372 8 application application NOUN iajs-3467 372 9 to to ADP iajs-3467 372 10 lifetime lifetime NOUN iajs-3467 372 11 data datum NOUN iajs-3467 372 12 . . PUNCT iajs-3467 373 1 thailand thailand PROPN iajs-3467 373 2 statistician statistician PROPN iajs-3467 373 3 . . PUNCT iajs-3467 374 1 2021 2021 NUM iajs-3467 374 2 , , PUNCT iajs-3467 374 3 19 19 NUM iajs-3467 374 4 , , PUNCT iajs-3467 374 5 484 484 NUM iajs-3467 374 6 - - SYM iajs-3467 374 7 500 500 NUM iajs-3467 374 8 . . PUNCT