Submit your paper : editorIJETjournal@gmail.com

Paper Title : Sarcasm Detection on Indonesian Politics Tweet Using Multi Labeling Method and Support Vector Machine

ISSN : 2395-1303
Year of Publication : 2020
10.29126/23951303/IJET-V6I3P15

Authors: Fariz Ardiansyah, Aghistina Kartikadewi

         


Abstract
- Detecting a sentence of sarcasm is considered as one of the difficult problems in sentiment analysis. In observing Twitter's social media in Indonesia, for some topics such as politics, people tend to criticize something using sarcasm. In previous studies, many researchers used the corpus of SentiWordNet or commonly called the Lexical method which has many limitations both from the number of words and from the language used that is only English and must be translated first into the desired language. In this study the author made Sarcasm Detection not with the help of a corpus or data dictionary but with the help of manual labeling using the Multi Labeling method. For the modeling method the author uses the TF-IDF method to convert words or sentences into vectors and Support Vector Machines. The whole modeling process uses the python programming language while for the display uses the PHP programming language. The training data set or training data for the author's model is obtained from the Twitter API crawling results with the help of a crawler engine made with the java programming language. Then for testing the author use the Confusion Matrix to determine the level of accuracy, sensitivity, specificity and precision of the proposed model. The final result of this study is for the SVM algorithm get an accuracy value of 86% while Naive Bayes as a comparison method gets an accuracy value of 78%. This proves that using machine learning that uses training data is proven to be able to overcome the limitations of the Lexical method because the training data can be added as much as the writer wants if the model is judged to be poor and the author does not need to do two classifications and there is no need to change the language process in the dictionary as in the Lexical method.

Reference
1. Kemp, S. (2018) Digital In 2018: World’s Internet Users Pass The 4 Billion Mark. Available at: https://wearesocial.com/blog/2018/01/global-digital-report-2018 (Accessed: 29 December 2018). 2. Lunando, E. and Purwarianti, A. (2013) ‘Indonesian Social Media Sentiment Analysis with Sarcasm Detection’, in 2013 International Conference on Advanced Computer Science and Information Systems, ICACSIS 2013, pp. 195–198. 3. Bamman, D. and Smith, N. A. (2015) ‘Contextualized Sarcasm Detection on Twitter’, in Proceedings of the Ninth International AAAI Conference on Web and Social Media, pp. 574–577. 4. Mohri, M., Rostamizadeh, A. and Talwalkar, A. (2018) Foundations of Machine Learning. 2nd edn. MIT Press. 5. Setiawan, E. (2012) Arti kata sarkasme - KamusBesar Bahasa Indonesia (KBBI) Online. Available at: https://kbbi.web.id/sarkasme (Accessed: 29 December 2018). 6. Khade, S. R. and Balwan, S. R. (2017) ‘Study and Analysis of Multi-Label Classification Methods in Data Mining’, International Journal of Computer Applications, 159(9), pp. 975–8887. 7. Wahid, D. H. and SN, A. (2016) ‘PeringkasanSentimenEsktraktif di Twitter Menggunakan Hybrid TF-IDF dan Cosine Similarity’, IJCCS (Indonesian Journal of Computing and Cybernetics Systems), 10(2), p. 207. 8. Sammut, C. and Webb, G. I. (2017) Encyclopedia of Machine Learning and Data Mining, Sammut& Webb. 2nd edn. Springer US. 9. Han, J., Kamber, M. and Pei, J. (2012) Data Mining – Concepts & Techniques. 3rd edn. Morgan Kaufmann Publishers.

Keywords
Sarcasm Detection, Multi Labelling, Support Vector Machine.