Sentiment Composition of Words with Opposing Polarities

Proceedings of NAACL-HLT 2016, pages 1102–1108, San Diego, California, June 12-17, 2016. c⃝2016 Association for Computational Linguistics Sentiment Composition of Words with Opposing Polarities Svetlana Kiritchenko and Saif M. Mohammad National Research Council Canada {svetlana.kiritchenko,saif.mohammad}@nrc-cnrc.gc.ca Abstract In this paper, we explore sentiment composi- tion in phrases that have at least one positive and at least one negative word—phrases like happy accident and best winter break. We compiled a dataset of such opposing polar- ity phrases and manually annotated them with real-valued scores of sentiment association. Using this dataset, we analyze the linguistic patterns present in opposing polarity phrases. Finally, we apply several unsupervised and su- pervised techniques of sentiment composition to determine their efficacy on this dataset. Our best system, which incorporates information from the phrase’s constituents, their parts of speech, their sentiment association scores, and their embedding vectors, obtains an accuracy of over 80% on the opposing polarity phrases. 1 Introduction The Principle of Compositionality states that the meaning of an expression is determined by the meaning of its constituents and by its grammatical structure (Montague, 1974). By extension, senti- ment composition is the determining of sentiment of a multi-word linguistic unit, such as a phrase or a sentence, based on its constituents. In this work, we study sentiment composition in phrases that in- clude at least one positive and at least one nega- tive word—for example, phrases such as happy ac- cident, couldn’t stop smiling, and lazy sundays. We refer to them as opposing polarity phrases. Such phrases present a particular challenge for automatic sentiment analysis systems that often rely on bag-of- word features. Word–sentiment associations are commonly cap- tured in sentiment lexicons. However, most existing manually created sentiment lexicons include only single words. Lexicons that include sentiment asso- ciations for multi-word phrases as well as their con- stituent words can be very useful in studying sen- timent composition. We refer to them as sentiment composition lexicons (SCLs). We created a sentiment composition lexicon for opposing polarity phrases and their constituent words (Kiritchenko and Mohammad, 2016c).1 Both phrases and single words were manually annotated with real-valued sentiment association scores using an annotation scheme known as Best–Worst Scal- ing.2 We refer to the created resource as the Sen- timent Composition Lexicon for Opposing Polarity Phrases (SCL-OPP). The lexicon includes entries for 265 trigrams, 311 bigrams, and 602 unigrams. In this paper, we use SCL-OPP to analyze regu- larities present in different kinds of opposing polar- ity phrases. We calculate the extent to which differ- ent part-of-speech combinations result in phrases of positive and negative polarity. We also show that for most phrases, knowing the parts of speech and po- larities of their constituents is not enough to reliably predict the sentiment of the phrase. We apply several unsupervised and supervised techniques of sentiment composition to determine their efficacy on predicting the sentiment of oppos- ing polarity phrases. Our experiments indicate that 1www.saifmohammad.com/WebPages/SCL.html#OPP 2Best–Worst Scaling has been shown to produce reliable real-valued sentiment association scores (Kiritchenko and Mo- hammad, 2016a). …

Content truncated to 500 words