Extending a sentiment lexicon with synonym–antonym datasets: SWNetTR++

Abstract

In our previous studies on developing a general-purpose Turkish sentiment lexicon, we constructed SWNetTRPLUS, a sentiment lexicon of 37K words. In this paper, we show how to use Turkish synonym and antonym word pairsto extend SWNetTR-PLUS by almost 33% to obtain SWNetTR++, a Turkish sentiment lexicon of 49K words. Theextension was done by transferring the problem into the graph domain, where nodes are words, and edges are synonym–antonym relations between words, and propagating the existing tone and polarity scores to the newly added wordsusing an algorithm we have developed. We tested the existing and new lexicons using a manually labeled Turkish newsmedia corpus of 500 news texts. The results show that our method yielded a significantly more accurate lexicon thanSWNetTR-PLUS, resulting in an accuracy increase from 72.2% to 80.4%. At this level, we have now maximized theaccuracy rates of translation-based sentiment analysis approaches, which first translate a Turkish text to English andthen do the analysis using English sentiment lexicons.