Toward Contextual Valence Shifters in Vietnamese Reviews

TOWARD CONTEXTUAL VALENCE SHIFTERS IN VIETNAMESE REVIEWS Thien Khai Tran1,2, Tuoi Thi Phan1 1 Faculty of Computer Science and Engineering, Ho Chi Minh City University of Technology VNU-HCM, Ho Chi Minh City, Vietnam 2 Faculty of Information Technology, Ho Chi Minh City University of Foreign Languages and Information Technology, Ho Chi Minh City, Vietnam 1,2thientk@cse.hcmut.edu.vn, 1tuoi@cse.hcmut.edu.vn Abstract. Valence shifters are complex linguistic structures that can modify the sentiment orientations of texts. In this paper, the authors concentrate on the study of shifters in Vietnamese texts and a discussion on the distribution of different types of shifters in the hotel reviews is presented. Finally, an approach for extracting the contextual valance shifters is proposed. Keywords: valance shifters, polarity shifters, sentiment shifter, Vietnamese shifters, sentiment analysis for Vietnamese, valance shifters of Vietnamese 1. Introduction Sentiment analysis assumes the task of identifying positive, negative, and neutral thoughts as well as emotional and subjective attitudes of the holder to the target audience, such as products, persons, or topics. Sentiment analysis is emerging as a research field attracting the scientific community. This includes the construction of emotional lexicons as a basis for other sentiment analysis problems at the document level, sentence level, and aspect level. This work can be done in many ways, with the simplest approach being to manually decide the polarities of sentiment words, and then to have a way of identifying the sentiment for each sentence or for the whole document based on the sentiment values of the words. However, this approach is not appropriate for subjectivity analysis because the sentiment values may be changed in context by the so-called valence shifters [1]. Valence shifters (or “polarity shifters”, “sentiment shifter”) are words (or phrases) that can change the sentiment orientations of texts. They are complex linguistic structures that may include explicit negations, contrasts, intensifiers, and diminishers, etc. [2]. In this paper, the authors present some approaches for contextual valence shifting detection of a Vietnamese sentiment analysis problem. We focus on rule-based methods that may be suitable for the complexity of the linguistic characteristics of Vietnamese. 152 The 2017 Conference on Computational Linguistics and Speech Processing ROCLING 2017, pp. 152-159  The Association for Computational Linguistics and Chinese Language Processing With this paper, the authors describe the research contributions: - Identifying many situations that cause valence shifters in Vietnamese texts. - Proposing an approach to deal with the problem of valence shifters of Vietnamese texts. To the best of our knowledge, this is the first work to investigate valence shifters in Vietnamese texts. The authors have organized the rest of this paper as follows: in Section 2 the related work is presented. In Section 3 the valence shifters for Vietnamese are introduced. In Section 4 the authors conclude the paper and discuss possibilities for future work. 2. Related Work The rise of the “contextual valence shifters” phenomenon has made traditional methods become ineffective when used to extract individual terms that indicate prior positive or negative polarity and build a set of emotional words. …

Content truncated to 500 words