The Journal of Society for e-Business Studies (한국전자거래학회지)

Society for e-Business Studies (한국전자거래학회)
권호 표지
DOI QR코드

DOI QR Code

Development and Validation of the Letter-unit based Korean Sentimental Analysis Model Using Convolution Neural Network

회선 신경망을 활용한 자모 단위 한국형 감성 분석 모델 개발 및 검증

  • Sung, Wonkyung (DTX Center, LG Electronics, Master, Graduate School of Information, Yonsei University) ;
  • An, Jaeyoung (Graduate School of Information, Yonsei University) ;
  • Lee, Choong C. (Graduate School of Information, Yonsei University)
  • Received : 2019.09.16
  • Accepted : 2020.01.21
  • Published : 2020.02.28
Download PDF
⟨ Previous Next ⟩

Abstract

This study proposes a Korean sentimental analysis algorithm that utilizes a letter-unit embedding and convolutional neural networks. Sentimental analysis is a natural language processing technique for subjective data analysis, such as a person's attitude, opinion, and propensity, as shown in the text. Recently, Korean sentimental analysis research has been steadily increased. However, it has failed to use a general-purpose sentimental dictionary and has built-up and used its own sentimental dictionary in each field. The problem with this phenomenon is that it does not conform to the characteristics of Korean. In this study, we have developed a model for analyzing emotions by producing syllable vectors based on the onset, peak, and coda, excluding morphology analysis during the emotional analysis procedure. As a result, we were able to minimize the problem of word learning and the problem of unregistered words, and the accuracy of the model was 88%. The model is less influenced by the unstructured nature of the input data and allows for polarized classification according to the context of the text. We hope that through this developed model will be easier for non-experts who wish to perform Korean sentimental analysis.

본 연구는 자모 단위의 임베딩과 회선 신경망을 활용한 한국어 감성 분석 알고리즘을 제안한다. 감성 분석은 텍스트에서 나타난 사람의 태도, 의견, 성향과 같은 주관적인 데이터 분석을 위한 자연어 처리 기술이다. 최근 한국어 감성 분석을 위한 연구는 꾸준히 증가하고 있지만, 범용 감성 사전을 사용하지 못하고 각 분야에서 자체적인 감성 사전을 구축하여 사용하고 있다. 이와 같은 현상의 문제는 한국어 특성에 맞지 않게 형태소 분석을 수행한다는 것이다. 따라서 본 연구에서는 감성 분석 절차 중 형태소 분석을 배제하고 초성, 중성, 종성을 기반으로 음절 벡터를 생성하여 감성 분석을 하는 모델을 개발하였다. 그 결과 단어 학습 문제와 미등록 단어의 문제점을 최소화할 수 있었고 모델의 정확도는 88% 나타내었다. 해당 모델은 입력 데이터의 비 정형성에 대한 영향을 적게 받으며, 텍스트의 맥락에 따른 극성 분류가 가능하게 되었다. 한국어 특성을 고려하여 개발된 본 모델이 한국어 감성 분석을 수행하고자 하는 비전문가에게 보다 쉽게 이용될 수 있기를 기대한다.

Keywords

References

  1. An, H. and I. Lee, “Extraction of Fashion Sensibility Vocabulary for Globalization,” Journal of Basic Design & Art, Vol. 14, No. 3, pp. 135-141, 2013.
  2. Baccianella, S., Esuli, A., and Sebastiani, F., "Sentiwordnet 3.0: An enhanced lexical resource for sentiment analysis and opinion mining," Lrec, Vol. 10, pp. 2200-2204, 2010.
  3. Bae, J. and Lee, C., “Sentiment Analysis with Skip-Connected LSTM,” Korea Information Science Society, Vol. 2017, No. 6, pp. 633-635, 2017.
  4. Bojanowski, P., Grave, E., Joulin, A., and Mikolov, T., "Enriching word vectors with subword information," Transactions of the Association for Computational Ling-uistics, Vol. 5, pp. 135-146, 2017. https://doi.org/10.1162/tacl_a_00051
  5. Bojanowski, P., Joulin, A., and Mikolov, T., "Alternative structures for character-level RNNs," arXiv preprint arXiv: 1511.06303, 2015.
  6. Chang, D., Kim, D., and Choi, Y., “Opinion Mining Based on Korean Phoneme Trigram-Signature,” The Korean Institute of Information Scientists and Engineers, Vol. 2015, No. 6, pp. 811-813, 2015.
  7. Choi, K. and Lee, C., “Sentiment analysis with GRU-Attention,” Korea Information Science Society, Vol. 2015, No. 12, pp. 557-559, 2015.
  8. Choi, S., "The modeling and training methods for syllable-based Korean word embeddings," Seoul National University, Master Dissertation, 2017.
  9. Choi, S., Lee, J., and Kwon, O., "A Morphological Analysis Method of Predicting Place-Event Performance by Online News Titles," The Journal of Society for e-Business Studies, Vol. 21, No. 1, pp. 15-32, 2016. https://doi.org/10.7838/jsebs.2016.21.1.015
  10. Chua, A. Y. and Banerjee, S., "Helpfulness of user-generated reviews as a function of review sentiment, product type and information quality," Computers in Human Behavior, Vol. 54, pp. 547-554, 2016. https://doi.org/10.1016/j.chb.2015.08.057
  11. Chung, J., Cho, K., and Bengio, Y., "A character-level decoder without explicit segmentation for neural machine translation," arXiv preprint arXiv:1603.06147, 2016.
  12. Collobert, R., Weston, J., Bottou, L., Karlen, M., Kavukcuoglu, K., and Kuksa, P., "Natural language processing (almost) from scratch," Journal of machine learning research, Vol. 12, pp. 2493-2537, 2011.
  13. Dos Santos, C. and Gatti, M., "Deep convolutional neural networks for sentiment analysis of short texts," Proceedings of Coling 2014, the 25th International Conference on Computational Linguistics: Technical Papers, pp. 69-78, 2014.
  14. Durfee, A., "Text Mining Promise and Reality," AMCIS 2006 Proceedings, p. 187, 2006.
  15. Eun, Z. and Park, S., “A Classification of Endings for an Efficient Morphological Analysis of Korean,” Journal of KIISE, Vol. 2000, No. 10, pp. 41-47, 2000.
  16. Harb, A., Plantie, M., Dray, G., Roche, M., Trousset, F., and Poncelet, P., "Web Opinion Mining: How to extract opinions from blogs?," in Proceedings of the 5th International Conference on Soft Computing as Transdisciplinary Science and Technology, pp. 211-217, 2008.
  17. Hong, T., Kim, E., and Cha, E., “The Prediction of dow jones and S&P500 index using SVM and news sentiment analysis,” The Journal of Internet Electronic Commerce Research, Vol. 17, No. 1, pp. 23-36, 2017.
  18. Hwang, H. and Lee, C., “Error correction in Korean morpheme recovery using deep learning,” Journal of KIISE, Vol. 42, No. 11, pp. 1452-1458, 2015. https://doi.org/10.5626/JOK.2015.42.11.1452
  19. Hwang, Y. and Choi, J., “The 21st century Sejong corpus properly-Using the Language Information Sharing Center,” National Institute of Korean Language, Vol. 26, No. 2, pp. 73-86, 2016.
  20. Kalchbrenner, N., Grefenstette, E., and Blunsom, P., "A convolutional neural network for modelling sentences," arXiv preprint arXiv:1404.2188, 2014.
  21. Kim, D., Kim, K., and Kim, J., "Character-based multi-category sentiment analysis on social media using deep learning algorithms," Korean Institute of Industrial Engineers, pp. 5082-5084, 2017.
  22. Kim, D., Park, J., and Choi, J., "A Comparative Study between Stock Price Prediction Models Using Sentiment Analysis and Machine Learning Based on SNS and News Articles," Journal of Information Technology Services, Vol. 13, pp. 211-233, 2014.
  23. Kim, G., Ock, K., and Lee, S., "Sentiment Dictionary Construction for Stock Fluctuation Prediction based on Security Company Reports," The Korean Institute of Information Scientists and Engineers, pp. 1022-1024, 2016.
  24. Kim, H. D. and Zhai, C., "Generating comparative summaries of contradictory opinions in text," in Proceedings of the 18th ACM conference on Information and Knowledge Management, pp. 385-394, 2009.
  25. Kim, J. O., Lee, S. S., and Yong, H. S., “Automatic Classification Scheme of Opinions Written in Korean,” Journal of KIISE: Databases, Vol. 38, No. 6, pp. 423-428, 2011.
  26. Kim, J. and Kim, D., "A Study on the Method for Extracting the Purpose-Specific Customized Information from Online Product Reviews based on Text Mining," The Journal of Society for e-Business Studies, Vol. 21, No. 2, pp. 151-161, 2016. https://doi.org/10.7838/jsebs.2016.21.2.151
  27. Kim, J., Oh, Y., and Chae, S., “The Construction of a Domain-Specific Sentiment Dictionary Using Graph-based Semi-supervised Learning Method,” Science of Emotion & Sensibility, Vol. 18, No. 1, pp. 103-110, 2015. https://doi.org/10.14695/KJSOS.2015.18.1.103
  28. Kim, M., Jang, H., Jo, Y., and Shin, H., “Korean Sentiment Analysis Corpus,” Korea Information Science Society, Vol. 2013, No. 6, pp. 650-652, 2013.
  29. Kim, M., Song, E., and Kim, Y., “A Design of Satisfaction Analysis System For Content Using Opinion Mining of Online Review Data,” Journal of Internet Computing and Services, Vol. 17, No. 3, pp. 107-113, 2016. https://doi.org/10.7472/jksii.2016.17.3.107
  30. Kim, S., Cho, H., and Kang, J., “The Status of Using Text Mining in Academic Research and Analysis Methods,” Journal of Information Technology and Architecture, Vol. 13, No. 2, pp. 317-329, 2016.
  31. Kim, S., Lee, Y. J., Shin, J., and Park, K. Y., “Text Mining for Economic Analysis,” BOK Economic Research Institute, Vol. 2019, No. 18, pp. 1-53, 2019.
  32. Kim, Y., "Convolutional neural networks for sentence classification," Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing, pp. 1746-1751, 2014.
  33. Kim, Y., Jernite, Y., Sontag, D., and Rush, A. M., "Character-aware neural language models," Thirtieth AAAI Conference on Artificial Intelligence, 2016.
  34. Kwon, S., "Sentiment Analysis of Movie Reviews using the Word2vec and RNN," Graduate School of Dongguk University, Master dissertation, 2017.
  35. Lankinen, M., Heikinheimo, H., Takala, P., Raiko, T., and Karhunen, J., "A character-word compositional neural language model for finnish," arXiv preprint arXiv:1612.03266, 2016.
  36. Lee, D. and Kim, K., “Web Site Keyword Selection Method by Considering Semantic Similarity Based on Word2Vec,” The Journal of Society for e-Business Studies, Vol. 23, No. 2, pp. 83-96, 2018. https://doi.org/10.7838/JSEBS.2018.23.2.083
  37. Lee, E., Kim, W., and Kim, S., "Korean Literature Dictionary," Korean Dictionary Research History, 1998.
  38. Lee, J., Lee, H., and Lee, H., “Research on Methods for Processing Nonstandard Korean Words on Social Network Services,” Journal of the Korea Society Industrial Information System, Vol. 21, No. 3, pp. 35-46, 2016.
  39. Lee, S., AI seeds Hangul corpus in 2007 and landed reason. 2016; Available from: https://www.bloter.net/archives/260569.
  40. Ling, W., et al., "Finding function in form: Compositional character models for open vocabulary word representation," arXiv preprint arXiv:1508.02096, 2015.
  41. Liu, B., “Sentiment analysis and opinion mining,” Synthesis Lectures on Human Language Technologies, Vol. 5, No. 1, pp. 1-167, 2012. https://doi.org/10.2200/S00416ED1V01Y201204HLT016
  42. Mu, J., Bhat, S., and Viswanath, P., "All-but-the-top: Simple and effective postprocessing for word representations," arXiv preprint arXiv:1702.01417, 2017.
  43. Park, E., "Supervised Feature Representations for Document Classification," Seoul National University, Doctoral dissertation, 2016.
  44. Park, S., “The media sentimental analysis of Yeongjong-do global MICE integrated resort,” International Journal of Tourism Management and Sciences, Vol. 31, No. 7, pp. 109-128, 2016. https://doi.org/10.21719/KJTR.31.7.6
  45. Peng, H., Cambria, E., and Zou, X., "Radical-based hierarchical embeddings for chinese sentiment analysis at sentence level," The Thirtieth International Flairs Conference, 2017.
  46. Rumelhart, D. E., Hinton, G. E., and Williams, R. J., "Learning representations by back-propagating errors," Cognitive modeling, Vol. 5, No. 3, p. 1, 1988.
  47. Salehan, M. and Kim, D. J., "Predicting the performance of online consumer reviews: A sentiment mining approach to big data analytics," Decision Support Systems, Vol. 81, pp. 30-40, 2016. https://doi.org/10.1016/j.dss.2015.10.006
  48. Santos, C. N. D. and Guimaraes, V., "Boosting named entity recognition with neural character embeddings," arXiv preprint arXiv:1505.05008, 2015.
  49. Sen, S. and Lerman, D., “Why are you telling me this? An examination into negative consumer reviews on the web,” Journal of interactive marketing, Vol. 21, No. 4, pp. 76-94, 2007. https://doi.org/10.1002/dir.20090
  50. Seo, J., Jo, H., and Choi, J., “Design for Opinion Dictionary of Emotion Applying Rules for Antonym of the Korean Grammar,” Journal of Korean Institute of Information Technology, Vol. 13, No. 2, pp. 109-117, 2015.
  51. Shin, H., Seo, M., and Byeon, H., “Korean Alphabet level Convolution Neural Network for Text Classification,” Korea Information Science Society, Vol. 2017, No. 6, pp. 587-589, 2017.
  52. Socher, R., Pennington, J., Huang, E. H., Ng, A. Y., and Manning, C. D., "Semi-supervised recursive autoencoders for predicting sentiment distributions," Proceedings of the conference on empirical methods in natural language processing, pp. 151-161, 2011.
  53. Song, J. and Lee, S., "Automatic Construction of Positive/Negative Feature-Predicate Dictionary for Polarity Classification of Product Reviews," Journal of KIISE: Software and Applications, Vol. 38, No. 3, pp. 157-168, 2011.
  54. Song, S. I., Lee, D. J., and Lee, S. G., "Identifying Sentiment Polarity of Korean Vocabulary Using PMI," Proceedings of the Korean Information Science Society Conference, pp. 260-265, 2010.
  55. Song, T., "Sentimatal Analysis on Food Safety Using Social Big Data," Korea Institute for Health and Social Affairs, Vol. 312, pp. 1-4, 2016.
  56. Wang, X., Liu, Y., Chengjie, S., Wang, B., and Wang, X., "Predicting polarities of tweets by composing word embeddings with long short-term memory," Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing, Vol. 1, pp. 1343-1353, 2015.
  57. Wu, X., "A density adjustment based particle swarm optimization learning algorithm for neural network design," 2011 International Conference on Electrical and Control Engineering, pp. 2829-2832, 2011.
  58. Yang, S. and Lee, C., “Sentiment Analysis using Latent Structural SVM,” Korea Information Science Society, Vol. 2015, No. 6, pp. 687-689, 2015.
  59. Yih, W. T., He, X., and Meek, C., "Semantic parsing for single-relation question answering," Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics, Vol. 2, pp. 643-648, 2014.
  60. Yune, H. J., Kim, H. J., and Chang, J. Y., “An efficient search method of product reviews using opinion mining techniques,” Journal of KIISE: Computing Practices and Letters, Vol. 16, No. 2, pp. 222-226, 2010.
  61. Zhang, X. and LeCun, Y., "Text understanding from scratch," arXiv preprint arXiv:1502.01710, 2015.
  62. Zheng, X., Chen, H., and Xu, T., "Deep learning for Chinese word segmentation and POS tagging," Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, pp. 647-657, 2013.