A Lexicon-based text classification model to analyse and predict sentiments from online reviews

The ubiquity of Internet has shown way to multitude of people to connect with each other beyond space and time. This has led to the colossal usage and hence popularity of online media as a platform to share opinions, exchange ideas, raise questions, show contempt and many such sorts of reviews. Thus web is transforming into a huge repository of textual data, which can be classified to fathom the sentiment and emotional state of an online user. There are many text classification algorithms proposed by different researchers that take in texts from online media and after text preprocessing, mining and classification, the sentiment of a user is predicted. In this paper, we are proposing a Lexicon-based text classification algorithm which is used to analyze and predict a user's sentiment polarity viz. positive, negative & neutral from online reviews. Our algorithm is different from other Lexicon-based algorithms in the context that it uses the three degrees of comparison viz. positive, comparative and superlative degrees on words; for each of the positive and negative sentiment words. Further, we have used the negation words to show how the accuracy of the system can be improved.