Enhancing Lexicon-Based Review Classification by Merging and Revising Sentiment Dictionaries

International Joint Conference on Natural Language Processing, pages 463–470, Nagoya, Japan, 14-18 October 2013. Enhancing Lexicon-Based Review Classification by Merging and Revising Sentiment Dictionaries Heeryon Cho Yonsei Institute of Convergence Technology, Yonsei University Incheon, Republic of Korea Jong-Seok Lee Songkuk Kim School of Integrated Technology Yonsei University Incheon, Republic of Korea {heeryon, jong-seok.lee, songkuk}@yonsei.ac.kr Abstract This paper presents a method of improving lexicon-based review classification by merg- ing multiple sentiment dictionaries, and selec- tively removing and switching the contents of merged dictionaries. First, we compare the positive/negative book review classification performance of eight individual sentiment dic- tionaries. Then, we select the seven dictionar- ies with greater than 50% accuracy and com- bine their results using (1) averaging, (2) weighted-averaging, and (3) majority voting. We show that the combined dictionaries per- form only slightly better than the best single dictionary (65.8%) achieving (1) 67.8%, (2) 67.7%, and (3) 68.3% respectively. To im- prove this, we combine seven dictionaries at a deeper level by merging the dictionary entry words and averaging the sentiment scores. Moreover, we leverage the skewed distribution of positive/negative threshold setting data to update the merged dictionary by selectively removing the dictionary entries that do not contribute to classification while switching the polarity of selected sentiment scores that hurts the classification performance. We show that the revised dictionary achieves 80.9% accura- cy and outperforms both the individual dic- tionaries and the shallow dictionary combina- tions in the book review classification task. 1 Introduction With the increase in opinion mining and senti- ment analysis-related researches, various lexical resources that define sentiment scores/categories have been constructed and made available. Ex- amples include SentiSense (de Albornoz et al., 2012), SentiWordNet (Baccianella et al., 2010), Micro-WNOp, and WordNet-Affect (Strapparava and Valitutti, 2004), which are based on a large English lexical database WordNet (Fellbaum, 1998), and AFINN (Nielsen, 2011), Opinion Lexicon (Hu and Liu, 2004), Subjectivity Lexi- con (Riloff and Wiebe, 2003) and General In- quirer (Stone and Hunt, 1963), which are manu- ally or semi-automatically constructed. These resources differ in their formats and sizes, but all can be utilized in the lexicon-based opinion min- ing and sentiment analysis. The increase in the number of sentiment re- sources naturally gives rise to two questions: (1) How are the performances of these resources different? (2) Can we construct a better senti- ment resource by combining and/or revising mul- tiple resources? We answer these questions by comparing the book review classification per- formance of single and combined sentiment re- sources, and present a simple ‘merge, remove, and switch’ approach that revises the entries of the sentiment resource to improve its classifica- tion performance. In the next section, we describe the experi- mental setup for evaluating the classification per- formance of sentiment resources. We then com- pare the positive/negative classification perfor- mance of eight widely known individual senti- ment resources in Section 3. Since individual sentiment resources are originally constructed in different formats, we standardize their formats. These standardized resources will be called sen- timent dictionaries or …

Content truncated to 500 words