Sentiment Recognition from Bangla Text

Abstract

Sentiment analysis is a very important area of the natural language processing. In general, sentiment classification means the analysis to determine the expression of a speaker whether he or she holds positive or negative opinion to a specific subject. With the rapid growth of e-commerce, sentiment analysis can greatly influence everyone in their real life. For example, product reviews on the Web have become an important source of information for customers’ decision making when they want to buy any product. As the reviews are often too many for customers to go through, how to automatically classify and detect the sentiment from them has become an important research problem. In this chapter, the authors present a Sentiment Analyzer that recognizes the Bangla sentiment or opinion about a subject from Bangla text. They construct some phrase patterns and calculate their sentiment orientation. They add tags to words in the Bangla text to construct the phrase pattern for positive and negative sentiment. Then the authors match the phrase pattern in Bangla text with their predefined phrase pattern and cumulate the sentiment orientation of each sentence.