Feature Extraction Method for Aspect-Based Sentiment Analysis

In our daily life, we take opinions of our friends and we are influenced in decision making process. Opinion is the view or the judgment about something. Opinion Mining (OM) or Sentiment Analysis (SA) is the computational analysis of public’s opinion, emotion, sentiments, and attitude toward entities and their attributes expressed in written text. These entities may be products, services, organizations, individuals, events, issues, or topics. In sentiment analysis, formal and informal opinion text like product reviews, news articles, tweets, forum discussions, blogs, and Facebook posts are also applicable to all domains. The main purpose of sentiment analysis is to extract the main opinions, on which the decision can be made very right. Paper intends to classify sentiment polarity on product review datasets by using Mutual Information as a feature selection method. Because product reviews are highly focused and they are opinion rich. After the feature selection, we aim to classify the extracted features with Naïve Bayes, SVM and Maximum Entropy to get the accurate sentiment polarity