28 Modelling Naïve Bayes for Tembang Macapat Classification Aji Prasetya Wibawa1, Yana Ningtyas1, Nimas Hadi Atmaja1, Ilham Ari Elbaith Zaeni1, Agung Bella Putra Utama1, Felix Andika Dwiyanto1, Andrew Nafalski2 1Universitas Negeri Malang, Indonesia 2University of South Australia, Australia Submitted: 2022-01-27. Revised: 2022-04-30. Accepted: 2022-06-02 Abstract The tembang macapat can be classified using its cultural concepts of guru lagu, guru wilangan, and guru gatra. People may face difficulties recognizing certain songs based on the established rules. This study aims to build classification models of tembang macapat using a simple yet powerful Naïve Bayes classifier. The Naive Bayes can generate high-accuracy values from sparse data. This study modifies the concept of Guru Lagu by retrieving the last vowel of each line. At the same time, guru wilangan’s guidelines are amended by counting the number of all characters (Model 2) rather than calculating the number of syllables (Model 1). The data source is serat wulangreh with 11 types of tembang macapat, namely maskumambang, mijil, sinom, durma, asmaradana, kinan- thi, pucung, gambuh, pangkur, dandhanggula, and megatruh. The k-fold cross-validation is used to evaluate the performance of 88 data. The result shows that the proposed Model 1 performs better than Model 2 in macapat classification. This promising method opens the potential of using a data mining classification engine as cultural teaching and preservation media. Keywords: Naïve Bayes, Classification, Tembang Macapat, Wulangreh How to Cite: Wibawa, A. P., Ningtyas, Y., Atmaja, N. H., Zaeni, I. A. E., Utama, A. B. P., Dwiyanto, F. A., & Nafalski, A. (2022). Modelling Naïve Bayes for Tembang Macapat Classification. Harmonia: Journal of Arts Research and Education, 22(1), 28-36 Harmonia: Journal of Arts Research and Education 22 (1) (2022), 28-36 Available online at http://journal.unnes.ac.id/nju/index.php/harmonia DOI: http://dx.doi.org/10.15294/harmonia.v22i1.34776 megatruh (Dewati, 2016; Setiyorini, 2016). Furthermore, tembang macapat is bound by the three main rules of guru lagu, guru gat- ra, and guru wilangan (Dwi Bambang Pu- tut Setiyadi, 2010). Guru lagu is the rhyme sound at the end of the word in each line. Guru Gatra is the number of lines in one verse, while Guru Wilangan is the number of syllables in each line (Novaeni, 2013). The creation of macapat has to follow these rules. A study reveals that macapat (e.g., se- kar pucung) has a very intellectual signifi- cance, but it also aims to train people to be INTRODUCTION Indonesia has various cultural diver- sity, including tembang. Tembang is tradi- tional Javanese poetry recited in song form (Daryanto, 2017), which was used to trans- mit long written texts of one form or anot- her in scenarios ranging from an evening in a palace to letters to a friend (Hatch, 1976). Macapat (Santosa, 2016) is one popu- lar genres of tembang. Generally, tembang Macapat have 11 types: maskumambang, mi- jil, sinom, durma, asmaradana, kinanthi, pu- cung, gambuh, pangkur, dandhanggula, and Corresponding author: E-mail: aji.prasetya.ft@um.ac.id p-ISSN 2541-1683|e-ISSN 2541-2426 Aji Prasetya Wibawa, et al., Modelling Naïve Bayes for Tembang Macapat Classification 29 religious and humanist (Saddhono & Pra- mestuti, 2018). The song incorporates po- sitive and negative Javanese educational terms (piwulang) of religion, ethics, morals, daily life, and government attitudes(D. B. P. Setiyadi, 2013). Serat Sabdajati, for example, uses macapat to explain Islamic doctrine and ethics (Fauziyyah, Warto, & Sariyatun, 2018). The positive should al- ways be followed, whereas the negative should be avoided. Furthermore, macapat can be used as a medium for character buil- ding (Agus et al., 2021). A bibliometric stu- dy of macapat from 1981 to 2021 found that most research is focused on cross-cultural and social change perspectives (Irmade & Winarto, 2021). None of them discuss this Javanese song from the computer science or technological aspect. The concept of guru wilangan, guru gatra, and guru lagu form patterns for computational modeling. The pattern dis- tinguishes one macapat text from others. Javanese artists, experts, or professional teachers may easily recognize these cultu- ral patterns. On the other hand, novice stu- dents may find it difficult to notice the tem- bang based on the pattern. Once the rules are broken, a song will not be identified. For example, we cannot label a tembang as pucung when it is less or more than four gatra with the following guru wilangan and guru lagu: 12u, 6a, 8i, 12a. The false tembang may confuse the singer, leading to failure in singing the song. Identifying the correct song will be increasingly difficult to do. This fatal error will be passed on to the next generation. This problematic situati- on prompted research on computational modeling of macapat classification. In data mining, classification is a typical technique for categorizing data points. This supervised machine learning approach has many classifiers. First, Lo- gistic regression is a probability-based model that works in only binary predicted variables, all independent predictors without missing values (Kasera & Joha- ri, 2021). Second, linear regression assu- mes that the independent and dependent variables are linearly connected; nonet- heless, it is susceptible to overfitting and noise (Schultz et al., 2021). On the other hand, the Support Vector Machine (SVM) displays training data in space, separated into categories by huge gaps; however, it delays generating probability estimates (Gaye, Zhang, & Wulamu, 2021). Final- ly, the Naive Bayes method assumes that all characteristics are independent of one another and contribute equally to the out- put (Mansour, Saleh, Badawy, & Ali, 2022). The Naive Bayes Classifier is a well-known generative algorithm that may use genera- tive algorithms to classify unknown data. These classification engines have excellent performance in classifying Javanese alpha- bets (Diqi & Muhdalifah, 2020; Handayani, Herwanto, Chandrika, & Arai, 2021; Rasyi- di, Bariyah, Riskajaya, & Septyani, 2021; Susanto, Atika Sari, Mulyono, & Doheir, 2021; Susanto, Sinaga, Sari, Rachmawanto, & Setiadi, 2018), vowels (Dewa & Afiaha- yati, 2018), speech levels (Ardhana, Cahya- ni, & Winarno, 2019; Nafalski & Wibawa, 2016), composition (Agus et al., 2021), dialect (Hidayatullah, Cahyaningtyas, & Pamungkas, 2020). None of research fo- cused on the macapat classification, which consider the gap as a novelty in the area of Javanese Culture classification. This research aims to create classifi- cation models of tembang macapat based on Naïve Bayes classifier. This Naive Bayes selection is based on its excellent classifica- tion performance when only using a small amount of training data (Gupta, Gupta, & Singh, 2005). Two models are proposed. The first model retains the cultural pat- terns: Guru Lagu, Guru Wilangan, and Guru Gatra. The second model assumes that not all users understand those cultural rules. Therefore, the second model limits the Guru Lagu by counting the number of vo- wels to replace the calculation of the num- ber of syllables. The second model repla- ces the Guru Wilangan by the probabilistic calculation of the characters in each gatra. This research is expected to facilitate the public understanding of the types of tem- bang macapat using the models. A promi- sing model will be decided after compa- Harmonia: Journal of Arts Research and Education 22 (1) (2022): 28-3630 ring the two proposed models. METHOD This data mining research is con- cerned with extracting usable and valuab- le information from enormous quantities of data. The suggested strategy allows for discovering intriguing patterns, which are deeply buried inside the data. The syste- matic procedures are described in the fol- lowing sections. Dataset and Preprocessing The data was collected from Serat Wulangreh book written by Kanjeng Su- suhunan Pakubuwana IV Surakarta Ha- diningrat (Wikandaru, Cathrin, Satria, & Rianita, 2020). This book has 11 types of Tembang which evaluated in this study, in- cluding maskumambang, durma, pucung, me- gatruh, gambuh, mijil, kinanthi, asmaradana, pangkur, sinom, dan dhandanggula. The se- lection of eight song for each type tembang, considering dandanggula as the song with the least number of stanzas. Thus, the total training data is 88 of tembang macapat. The next stage is preprocessing, an initial activity required before the classifi- cation process (Andini, 2013). This process produces a variable from the tembang data- set containing a value set representing the information contained in the tembang. This stage splits each row and calculates the number of lines in the lyric. The following process is to generate model 1, following the rule of guru lagu, by removing space and consonants in each line. The remaining vowel letter on each line will be calculated to get the guru wilan- gan. The last vowel is recognized as guru lagu. Table 1 presents the preprocessing re- sult of model 1. Model 2 modifies the rule of guru wilangan. As in model 1, this modification starts by removing the space, followed by calculating the characters of each line. The calculation is recognized as the modified guru wilangan. The modification considers the reality that mistakes (e.g., typographi- cal errors) could have happened during the Tembang writing. Determining the last vowel letter is completed by displaying only vowel letters on each line and then selecting the last vowel letter. Table 2 pre- sents the preprocessing result of the guru lagu rule in model 2. The table is similar to Table 1 except for the result. In the same song, the result of table 2 is higher than Table 1 due to all characters’ recognition. Table 1. Pre-processing Result of Model 1 Line Lyric Vowel Result 1 Sekar gambuh ping catur e a a u i a u 7u 2 Kang cinatur polah kang kalantur a i a u o a a a a u 10u 3 Tanpa tutur katula-tula katali a a u u a u a u a a a i 12i 4 Kadaluwarsa katutuh a a u a a a u u 8u 5 Kepatuh pan dadi awon e a u a a i a o 80 Table 2. Pre-processing Result of Model 2 Line Lyric Vowel Result 1 Sekar gambuh ping catur e a a u i a u 20u 2 Kang cinatur polah kang kalantur a i a u o a a a a u 27u 3 Tanpa tutur katula-tula katali a a u u a u a u a a a i 27i 4 Kadaluwarsa katutuh a a u a a a u u 18u 5 Kepatuh pan dadi awon e a u a a i a o 18o Naïve Bayes Classification This study uses the Naïve Bayes clas- sification approach. The simple probabili- ty classification calculates a set of proba- bilities, summing up frequency and value combination from a dataset (Suardani, Bhaskara, & Sudarma, 2018). Furthermore, an attribute or variable in Naïve Bayes is assumed to be independent or have no re- lation to each other (Manino, Tran-Thanh, & Jennings, 2019). One of Naïve Bayes advantages is its robust performance for classifying limited training data. This advantage is essential and suitable for this study. Equation 1 pre- sents the method of Naïve Bayes (Assiroj, 2018). P(H│X)=P(X│H)P(X) (1) Aji Prasetya Wibawa, et al., Modelling Naïve Bayes for Tembang Macapat Classification 31 Where: X = data with unknown class H = hypothesis from data X P(H│X) = the probability of hypothesis H based on condition X P(X│H) = the probability of hypothesis X based on condition H P(X) = probability of X Evaluation A confusion matrix is used to evalu- ate the classification results. It is a matrix table-based method for determining the accuracy of data generated by an algorithm (Haghighi, Jasemi, Hessabi, & Zolanvari, 2018). Table 3 contains the matrix table for the dataset with two classes. Table 3. Confusion Matrix Actual Predictive Positive Negative Positive True Positive False Negative Negative False Positive True Negative If the classification results are pre- cisely classified into the correct class, the classification results are true positives. False-negative values are returned if incor- rectly classified the classification results into the correct class. False positives are calculated if the classification results are precisely categorized in the incorrect class. True-negative values are returned if the classification results are incorrectly classi- fied into the incorrect class. The confusion matrix will generate values for accuracy, precision, and recall (Andriani, 2012). Ac- curacy, precision, and recall can be calcula- ted using equations 2 to 4. Accuracy= TP+TN x 100% (2) P+TN+FP+FN Precision= TP x 100% (3) TP+FP Recall= TP x 100% (4) TP+FN RESULT AND DISCUSSION The classification technique, the Naïve Bayes approach, tries to classify the class or type of tembang macapat. The clas- sification performance is evaluated using k-fold cross-validation, which generates a confusion matrix table. K-fold cross-vali- dation divides the data into k folds, whe- re k is the fold value. Each fold’s test data was derived from the previous k-1 folds categorization. Four-fold k-fold cross-vali- dation will be used in this study. Table 4. The class of Tembang Macapat Class Type Characteristic Gatra a Maskumambang 12i, 6a, 8i, 8a 4 b Durma 12a, 7i, 7a, 7a, 8i, 5a, 7i 7 c Pucung 12u, 6a, 8i, 12a 4 d Megatruh 12u, 8i, 8u, 8i, 8o 5 e Gambuh 7u, 10u, 12i, 8u, 8o 5 f Mijil 10i, 6o, 10e, 10i, 6i, 6u 6 g Kinanthi 8u, 8i, 8a, 8i, 8a, 8i 6 h Asmaradana 8i, 8a, 8e, 8a, 7a, 8u, 8a 7 i Pangkur 8a, 11i, 8u, 7a, 12u, 8a, 8i 7 j Sinom 8a, 8i, 8a, 8i, 7i, 8u, 7a, 8i, 12a 9 k Dhandanggula 10i, 10a, 8e, 7u, 9i, 7a, 6u, 8a, 12i, 7a 10 In this study, both Model 1 and Mo- del 2 are tested using four-fold values. The 88 tembang macapat were separated into four-folds, resulting in an equal distributi- on of 22 songs. We use 22 test data and 66 training data in each fold. As illustrated in Table 4, each existing tembang macapat class is denoted by a variable alphabet let- ter from a to k. Table 5 contains the confusion matrix for the original tembang macapat (Model 1), with k=4. The table demonstrates that all eleven tembang of each class are accurate- ly categorized. In other words, the charac- teristics of tembang (Table 4) are recogni- zed by Model 1. Harmonia: Journal of Arts Research and Education 22 (1) (2022): 28-3632 Table 5. Confusion Matrix Modifikasi Model 1 True class Classification Results a b c d e f g h i j k a 8 0 0 0 0 0 0 0 0 0 0 b 0 8 0 0 0 0 0 0 0 0 0 c 0 0 8 0 0 0 0 0 0 0 0 d 0 0 0 8 0 0 0 0 0 0 0 e 0 0 0 0 8 0 0 0 0 0 0 f 0 0 0 0 0 8 0 0 0 0 0 g 0 0 0 0 0 0 8 0 0 0 0 h 0 0 0 0 0 0 0 8 0 0 0 i 0 0 0 0 0 0 0 0 8 0 0 j 0 0 0 0 0 0 0 0 0 8 0 k 0 0 0 0 0 0 0 0 0 0 8 The confusion matrix from Model 2 is presented in Table 6. Table 6. Confusion Matrix Modifikasi Model 2 True class Classification results a b c d e f g h i j k a 5 0 3 0 0 0 0 0 0 0 0 b 0 7 0 0 0 0 0 0 1 0 0 c 2 0 6 0 0 0 0 0 0 0 0 d 0 0 2 6 0 0 0 0 0 0 0 e 0 0 0 1 7 0 0 0 0 0 0 f 0 0 0 0 0 8 0 0 0 0 0 g 0 0 0 0 0 0 8 0 0 0 0 h 0 0 0 0 0 1 0 7 0 0 0 i 0 0 0 0 0 0 0 0 8 0 0 j 0 0 0 0 0 0 2 0 0 6 0 k 0 0 0 0 0 0 0 1 1 0 6 According to Table 6, eight tembang macapat are associated with the maskumam- bang (a) class. The Naïve Bayes technique classified five of the eight songs correctly. While the remaining three were incor- rectly labeled as pucung due to the pattern similarity. Furthermore, the Naïve Bayes algorithm can accurately identify seven of eight durma (b) songs. Only one durma is miscategorized as pangkur, because of the equal number of gatra. We tested eight pucung (c) tembang and found that two were mislabelled as maskumambang (a). Six of eight megatruh (d) is well classified, while the rest is cate- gorized as pucung. The Naïve Bayes approach accurately classifies seven of eight Gambuh (e) songs. An error is confirmed since one song was labeled megatruh (d). On the mijil (f), ki- nanthi (g), and pangkur (i), the Naïve Bayes algorithm correctly classifies all instances. Seven Asmaradana (h) songs were correctly classified, while the last remai- ning song is labeled as Mijil (f). The Naive Bayes approach successfully classified six songs as sinom (j) class, while two data are categorized as Kinanthi (g). Finally, eight dhandanggula (k) were categorized as six correct classifications and two incorrect songs. The two mistakes are labeled as as- maradana (h) and pangkur (i), respectively. The classification performances of the two proposed models can be seen in Table 7. Table 7. Classification Performance of Devel- oped models Model Accuracy Precision Recall 1 100% 100% 100% 2 84.09% 86.63% 84.09% Based on Table 7, Model 1 has bet- ter results when compared to Model 2, which has a difference in the accuracy of 15.91%. Model 1 obtained 100% accuracy, precision, and recall. For Model 2, the ob- tained accuracy and recall are 84.09%, with 86.63% precision. The outstanding results show that the proposed model that accom- modates the original rule of guru lagu and guru wilangan is much better than model 2. While the tembang follows the rules, it can easily recognize by model 1. Model 2 could help predict the tembang created by a novi- ce, who does not have enough understan- ding of the cultural rules. The proposed models show that Naïve Bayes can classify the macapat cor- rectly. The models can adopt the cultural rules (guru lagu and guru wilangan) into numeric features as inputs of the classifi- cation engine. In other words, the models are ready to use for tembang macapat clas- sification. A promising usage of the developed models is for Javanese cultural learning Aji Prasetya Wibawa, et al., Modelling Naïve Bayes for Tembang Macapat Classification 33 and preservation. The tembang macapat is still mentioned as teaching material in the Indonesian 2013 curriculum (D.B.P. Se- tiyadi & Haryono, 2018). Macapat can be used as a vocabulary learning media (Wa- hyudi, 2017). On the other hand, macapat is a suitable medium for early character education (Suciptaningsih, Widodo, & Ha- ryati, 2017). In order to teach the tembang macapat, teachers have enough experience and knowledge about titilaras (scale) and the meaning of tembang (Pairin M. Basir & Marifatulloh, 2018). The teachers’ compe- tence, willingness, and motivation impact the effectiveness of teaching tembang. The developed classification models as the of Information and Communication Technologies (ICTs) representation could boost the teachers’ performance. Teachers are adjusting their digital skills, an ongoing process that must be ongoing and in which knowledge gaps continue to undermine its deployment (Sánchez Prieto, Trujillo Tor- res, Gómez García, & Gómez García, 2020). Technology can create chances for motiva- ted learners. However, it is unlikely to re- sult in motivation or independent conduct in most students (Stockwell & Reinders, 2019). In other words, a bond between te- achers and students is badly required to fill the gap in technology usage. CONCLUSIONS This study succeeded in developing a macapat song classification model using Naïve Bayes. Model 1 keeps the original cultural characteristics of Tembang macapat, which outperforms Model 2, which com- putationally modified the guru wilangan. For further research, it is recommended to add more data or use other classification methods to compare the performance of developed models. A broader view from text mining and machine translation fields could be beneficial for understanding the textual meaning of macapat. Finally, there is no reason for instructors to neglect crea- tivity and innovation in their teaching met- hod, including implementing these classi- fication models. Once teachers recognize the importance of Tembang as a worthy cultural asset and a medium of education, the Javanese culture preservation could be achieved as the growth of ICT. REFERENCES Agus, C., Saktimulya, S. R., Dwiarso, P., Widodo, B., Rochmiyati, S., & Dar- mowiyono, M. (2021). Revitaliza- tion of local traditional culture for sustainable development of national character building in Indonesia. In Innovations and Traditions for Sus- tainable Development (pp. 347–369). https://doi.org/10.1007/978-3-030- 78825-4_21 Andini, S. (2013). Klasifikasi dokumen teks menggunakan algoritma Naive Bayes dengan bahasa pemprogram- an Java. Teknologi Informasi & Pendi- dikan, 6(2), 140–147. Andriani, A. (2012). Penerapan algoritma C4.5 pada program klasifikasi ma- hasiswa dropout. Seminar Nasional Matematika 2012, 139–147. Ardhana, A. P., Cahyani, D. E., & Wi- narno. (2019). Classification of Java- nese language level on articles us- ing multinomial Naive Bayes and N-Gram methods. Journal of Physics: Conference Series, 1306(1), 012049. h t t p s : / / d o i . o r g / 1 0 . 1 0 8 8 / 1 7 4 2 - 6596/1306/1/012049 Assiroj, P. (2018). Kajian perbandingan teknik klasifikasi algoritma C4.5, Naïve Bayes dan Cart untuk prediksi kelulusan mahasiswa (Studi kasus: STMIK Rosma Karawang). Media Informatika, 15(2), 1–17. https://doi. org/10.5281/zenodo.1184054 Daryanto, J. (2017). Pendidikan karak- ter dalam Serat Sanasunu karya R. Ng. Yasadipura II. Jurnal Pendidikan Dasar, 5(2), 77–81. Harmonia: Journal of Arts Research and Education 22 (1) (2022): 28-3634 Dewa, C. K., & Afiahayati. (2018). Suitable CNN weight initialization and acti- vation function for Javanese vow- els classification. Procedia Computer Science, 144, 124–132. https://doi. org/10.1016/j.procs.2018.10.512 Dewati, U. (2016). Upaya ngundhakake kawasisan nulis tembang Pucung lu- mantar medhia gambar siswa kelas IXA SMPN 6 Trenggalek taun ajaran 2015/2016. BARADHA, 1(1), 0–13. Diqi, M., & Muhdalifah, M. (2020). Design and building Javanese script classi- fication in the state museum of So- nobudoyo Yogyakarta. International Journal of Informatics and Computation, 1(2), 35. https://doi.org/10.35842/ ijicom.v1i2.18 Fauziyyah, F. I., Warto, W., & Sariyatun, S. (2018). Ronggowarsito’s conceptt of Islamic heosophy in serat sabdajati. International Journal of Multicultural and Multireligious Understanding, 5(2), 177. https://doi.org/10.18415/ ijmmu.v5i2.237 Gaye, B., Zhang, D., & Wulamu, A. (2021). Improvement of Support Vector Machine algorithm in big data back- ground. Mathematical Problems in Engineering, 2021, 1–9. https://doi. org/10.1155/2021/5594899 Gupta, A., Gupta, S., & Singh, D. (2005). A systematic review of classifica- tion techniques and implementation of ID3 Decision Tree Algorithm. 4th International Conference on System Modeling & Advancement in Research Trends (SMART), 144–152. Haghighi, S., Jasemi, M., Hessabi, S., & Zo- lanvari, A. (2018). PyCM: Multiclass confusion matrix library in Python. Journal of Open Source Software, 3(25), 729. https://doi.org/10.21105/ joss.00729 Handayani, A. N., Herwanto, H. W., Chandrika, K. L., & Arai, K. (2021). Recognition of handwritten Javanese script using backpropagation with zoning feature extraction. Knowledge Engineering and Data Science, 4(2), 117. https://doi.org/10.17977/um- 018v4i22021p117-127 Hatch, M. (1976). The song is ended: Changes on the use of Macapat in Central Java. Asian Music, 7(2), 59. https://doi.org/10.2307/833789 Hidayatullah, A. F., Cahyaningtyas, S., & Pamungkas, R. D. (2020). Attention- based CNN-BiLSTM for dialect iden- tification on Javanese text. Kinetik: Game Technology, Information System, Computer Network, Computing, Elec- tronics, and Control, 5(4), 317–324. https://doi.org/10.22219/kinetik. v5i4.1121 Irmade, O., & Winarto, W. (2021). Visual- isasi penelitian tentang tembang Ma- capat dari tahun 1981 – 2021 : Anali- sis Bibliometrik. Gelar : Jurnal Seni Budaya, 19(1), 97–104. https://doi. org/10.33153/glr.v19i1.3789 Kasera, M., & Johari, R. (2021). Prediction using machine learning in sports: a case study. In Lecture Notes on Data Engineering and Communications Technologies, vol 54. (pp. 805–813). https://doi.org/10.1007/978-981- 15-8335-3_61 Manino, E., Tran-Thanh, L., & Jennings, N. R. (2019). On the efficiency of data collection for multiple Naïve Bayes classifiers. Artificial Intelligence, 275, 356–378. https://doi.org/10.1016/j. artint.2019.06.010 Mansour, N. A., Saleh, A. I., Badawy, M., & Ali, H. A. (2022). Accurate detection of Covid-19 patients based on Fea- ture Correlated Naïve Bayes (FCNB) classification strategy. Journal of Ambient Intelligence and Humanized Computing, 13(1), 41–73. https://doi. org/10.1007/s12652-020-02883-2 Nafalski, A., & Wibawa, A. P. (2016). Machine translation with Java- nese speech levels’ classification. Informatics, Control, Measurement in Economy and Environment Pro- tection, 6(1), 21–25. https://doi. org/10.5604/20830157.1194260 Novaeni, G. A. (2013). Menggali konsep Tri Hita Karana dalam geguritan Sekar Mulat sebagai upaya pemben- Aji Prasetya Wibawa, et al., Modelling Naïve Bayes for Tembang Macapat Classification 35 tukan karakter bangsa. Jumantara, Vol. 4, pp. 113–124. Pairin M. Basir, U., & Marifatulloh, S. (2018). The Art of Tembang Macapat: Exclusiveness of the forms, value aspects, and learning approach. 222(SoSHEC), 226–230. https://doi.org/10.2991/ soshec-18.2018.49 Rasyidi, M. A., Bariyah, T., Riskajaya, Y. I., & Septyani, A. D. (2021). Classifica- tion of handwritten Javanese script using random forest algorithm. Bul- letin of Electrical Engineering and In- formatics, 10(3), 1308–1315. https:// doi.org/10.11591/eei.v10i3.3036 Saddhono, K., & Pramestuti, D. (2018). Sekar macapat pocung: study of religious values based on the lo- cal wisdom of javanese culture. EL HARAKAH (TERAKREDITASI), 20(1), 15. https://doi.org/10.18860/ el.v20i1.4724 Sánchez Prieto, J., Trujillo Torres, J. M., Gómez García, M., & Gómez Gar- cía, G. (2020). Gender and digital teaching competence in dual voca- tional education and training. Educa- tion Sciences, 10(3), 84. https://doi. org/10.3390/educsci10030084 Santosa, P. (2016). Fungsi sosial kema- syarakatan tembang Macapat (Com- munity Social Functions of Macapat). Widyaparwa, 44(2), 97–109. https:// doi.org/10.26499/wdprw.v44i2.138 Schultz, B. G., Joukhadar, Z., Nattala, U., Quiroga, M. del M., Bolk, F., & Vo- gel, A. P. (2021). Best practices for supervised machine learning when examining biomarkers in clinical populations. In Big Data in Psychia- try & Neurology (pp. 1–34). https:// doi.org/10.1016/B978-0-12-822884- 5.00013-1 Setiyadi, D. B. P. (2013). Discourse analysis of serat Kalatidha: Javanese cogni- tion system and local wisdom. Asian Journal of Social Sciences & Humani- ties, 2(4), 292–300. Setiyadi, D.B.P., & Haryono, P. (2018). Tembang macapat lyrics-based char- acter education learning materials for secondary school students. In Character Education for 21st Century Global Citizen (1st ed., pp. 299–306). Routledge. Setiyadi, Dwi Bambang Putut. (2010). Wacana tembang Macapat sebagai pengungkap sistem kognisi dan ke- arifan lokal Etnik Jawa. Kajian Lin- guistik Dan Sastra, 22 No.2(1), 193– 210. https://doi.org/10.23917/kls. v22i2.4375 Setiyorini, A. (2016). Aplikasi tembang Macapat berbasis multimedia. Jurnal Informasi Interaktif, 1(2), 87–92. Stockwell, G., & Reinders, H. (2019). Tech- nology, motivation and autonomy, and teacher psychology in language learning: exploring the myths and possibilities. Annual Review of Ap- plied Linguistics, 39, 40–51. https:// doi.org/10.1017/S0267190519000084 Suardani, L. G. P., Bhaskara, I. M. A., & Sudarma, M. (2018). Optimization of feature selection using Genetic Algo- rithm with Naïve Bayes classification for home improvement recipients. International Journal of Engineering and Emerging Technology, 3(1), 66–70. Suciptaningsih, O. A., Widodo, S., & Haryati, T. (2017). Teaching character education to primary school students through Javanese ethnolinguistics. 118, 747–756. https://doi.org/10.2991/ icset-17.2017.124 Susanto, A., Atika Sari, C., Mulyono, I. U. W., & Doheir, M. (2021). Histogram of gradient in K-Nearest Neighbor for Javanese alphabet classification. Scientific Journal of Informatics, 8(2), 289–296. https://doi.org/10.15294/ sji.v8i2.30788 Susanto, A., Sinaga, D., Sari, C. A., Rach- mawanto, E. H., & Setiadi, D. R. I. M. (2018). A high performace of local binary pattern on classify Javanese character classification. Scientific Journal of Informatics, 5(1), 8. https:// doi.org/10.15294/sji.v5i1.14017 Wahyudi, T. T. (2017). Macapat as a model and method of learning English-Java- nese vocabulary. KnE Social Sciences, Harmonia: Journal of Arts Research and Education 22 (1) (2022): 28-3636 1(3), 311. https://doi.org/10.18502/ kss.v1i3.751 Wikandaru, R., Cathrin, S., Satria, E., & Rianita, D. (2020). Critical analysis of Javanese epistemology and its rel- evance to science development in In- donesia. Jurnal Humaniora, 32(3), 206. https://doi.org/10.22146/jh.49065