This document discusses sentiment analysis, which is the computational study of opinions expressed in text. It defines sentiment analysis as identifying the positive, negative, or neutral orientation of opinions expressed in documents, sentences, or features of an object. The document outlines that sentiment analysis can be performed at the word, sentence, or document level. It also explains that sentiment analysis aims to structure unstructured text by discovering quintuples that represent opinions in terms of the object, feature, sentiment, opinion holder, and time. The document provides examples of applications of sentiment analysis like review classification and product feature analysis.