Sentiment analysis of movie reviews using lexicon approach

Sentiment analysis is a language processing task which is used to find out opinion expressed in online reviews to categorize it into different classes like positive, negative or neutral. The paper aims to summarize the movie reviews at aspect level so that user can easily find out which aspects of movie are liked and disliked by user. Before finding aspect and its respective opinion of movie, proposed system performs subjectivity analysis. Subjectivity analysis is one of the important and useful tasks in sentiment analysis. Online reviews may consist of both objective and subjective sentences. Among these, objective sentences consist of only factual information and no sentiments or opinion. Hence subjective sentences are considered for further processing i.e. to find feature-opinion pair and to find summery at aspect level. In this paper, two different methods are implemented for finding subjectivity of sentences and then rule based system is used to find feature-opinion pair and finally the orientation of extracted opinion is revealed using two different method. Initially the proposed system uses SentiWordNet approach to find out orientation of extracted opinion and then it uses the method which is based on lexicon consisting list of positive and negative words.