Sentiment Analysis for Social Text

This chapter is focusing on sentiment analysis—the computational detection and study of opinions and viewpoints underlying a text span-in social text settings: short, informal, and noisy text spans. The types of materials we bring here are not only effective in addressing various of sentiment tasks, but they are effectively teaching and guiding us during the whole reading as well. This chapter was written as a response to the great interest that sentiment analysis attracts, which has been growing in parallel with the proliferation of digital communication and the development of artificial intelligence practices. The methods, applications, and resources that we bring here can grant access into the thoughts, opinions, and beliefs of populations—a vast reservoir of information that was previously unobtainable, and it enables researchers to provide significant insights into opinions without having to directly survey populations, a time-consuming and expensive task. The chapter contains a short ontology of the field and explores the relevant tasks for social data: lexical-, aspect-, and sentence-level sentiment analysis.