Visualizing Sentiments and Emotions
Nan Cao,Weiwei Cui
DOI: https://doi.org/10.2991/978-94-6239-186-4_6
2016-01-01
Abstract:Sentiment analysis, also known as opinion mining, is one of the most important text mining tasks and has been widely used for analyzing, for example, reviews or social media data for various of applications, including marketing and customer service. In general, “sentiment analysis aims to determine the attitude of a speaker or a writer with respect to some topic or the overall contextual polarity of a document. The attitude may be his or her judgment or evaluation, affective state (i.e., the emotional state of the author when writing), or the intended emotional communication (i.e., the emotional effect the author wishes to have on the reader)” (Wikipedia, Sentiment analysis—Wikipedia, the free encyclopedia, 2006. https://en.wikipedia.org/wiki/Sentiment_analysis [17]). The sentiment analysis result is usually a score that ranges from −1 to 1 (after normalization) with −1 indicating the most negative, 1 indicating the most positive, and 0 indicating neutral. This sentiment score is usually treated as an attribute of the corresponding text, which can be intuitively differentiated by colors that range from, for example, red (−1) to green (1). In this chapter, we introduce the state-of-the-art sentiments visualization techniques that can be largely classified into two categories including (1) the techniques for summarizing the sentiment dynamics over time, and (2) the techniques for assisting sentiment analysis.