스탠포드 감성 트리 말뭉치를 이용한 감성 분류 시스템
Sentiment Analysis System Using Stanford Sentiment Treebank
0
조회
0
다운로드
1
인용
0
참고문헌
키워드
초록
본 연구는 스탠포드 감성 트리 말뭉치를 이용하여 감성 분류 시스템을 구현하였으며, 분류기로는 지지벡터기계(Support Vector Machines)를 이용하여 긍정, 중립, 부정 등의 3가지 감성으로 분류하였다. 먼저 감성 문장의 품사를 부착한 후 의존구조를 부착하였다. 트리 말뭉치의 모든 노드와 감성 태그를 자동으로 추출하여 문장 레벨의 지지벡터 분류시스템과 노드 레벨의 지지벡터 분류 시스템을 각각 구현하였다. 자질로는 어휘, 품사, 감성어휘, 의존관계, 형제관계 등다양한 자질의 조합을 이용하였다. 평가 말뭉치를 이용하여 3클래스로 분류한 결과, 노드 단위에서는 74.2%, 문장 단위에서는 67.0%의 정확도를 얻었으나 2클래스 분류에서는 현재 알려진 최고의 시스템에 어느 정도 필적하는 성능을 거두었다.
Abstract
The main goal of this research is to build a sentiment analysis system which automatically determines user opinionsof the Stanford Sentiment Treebank in terms of three sentiments such as positive, negative, and neutral. Firstly, sentiment sentencesare POS tagged and parsed to dependency structures. All nodes of the Treebank and their polarities are automaticallyextracted from the Treebank. We train two Support Vector Machines models. One is for a node level classification and theother is for a sentence level. We have tried various type of features such as word lexicons, POS tags, Sentiment lexicons,head-modifier relations, and sibling relations. Though we acquired 74.2% in accuracy on the test set for 3 class node levelclassification and 67.0% for 3 class sentence level classification, our experimental results for 2 class classification are comparableto those of the state of art system using the same corpus.
참고문헌 (13)
- 이공주, "Compositional rules of Korean auxiliary predicates for sentiment analysis", 한국마린엔지니어링학회지, 37(3), pp.291~299 (2013), 한국마린엔지니어링학회[DOI]
- B. Liu, "Opinion observer :Analyzing and comparing opinions on the web", Proceedings of the 14th international World Wide Web conference, pp.342~451 (2005)
- A. M. Popescu, "Extracting product features and opinions from reviews", Proceedings of Conference on Empirical Methods on Natural Language Processing, pp.339~346 (2005)
- P. D. Turney, "Thumbs up or thumbs down? Semantic orientation applied to unsupervised classification of reviews", Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics, pp.417~424 (2002)
- R. Socher, "Recursive deep models for semantic compositionality over a sentiment treebank", Proceedings of Conference on Empirical Methods on Natural Language Processing (2013)
- J. Bollen, "Twitter mood predicts the stock market", Journal of Computational Science, 2(1), pp.1~8 (2011)
- S. Asur, "Predicting the future with social media", Proceedings of the 2010IEEE, ACM International Conference on Web Intelligence and Intelligent Agent Technology, pp.2010, 1[DOI]
- A. Pak, "Twitter as a corpus for sentiment analysis and opinion mining", Proceedings of the Seventh Conference on International Language Resources and Evaluation, pp.1320~1326 (2010)
- Twitter sentiment classification using distant supervision, "A. Go", Stanford University, ~, pp.Technical report CS224N (2009)
- B. O'Connor, "From tweets to polls : Linking text sentiment to public opinion time series", Proceedings of the Fourth International AAAI Conference on Weblogs and Social Media, pp.122~129 (2010)
- C. D. Manning, "The stanford coreNLP natural language processing toolkit", Proceedings of 52nd Annual Meeting of the Association for Computational Linguistics : System Demonstrations, pp.55~60 (2014)
- M. Ganapathibhotla, "Mining opinions in comparative sentences", Proceedings of the 22nd International Conference on Computational Linguistics, pp.18~22 (2008)
- C. C. Chang, "LIBSVM : a library for support vector machines", ACM Transactions on Intelligent Systems and Technology, 2(2), pp.27:1~27:27 (2011)