A Novel Hybrid Approach Based on Maximum Entropy Classifier for Sentiment Analysis of Malayalam Movie Reviews

Anagha M; Raveena R Kumar; Sreetha K; P C Reghu Raj

Volume : IV, Issue : VI, June - 2015

A Novel Hybrid Approach Based on Maximum Entropy Classifier for Sentiment Analysis of Malayalam Movie Reviews

Anagha M, Raveena R Kumar, Sreetha K, P C Reghu Raj

Abstract :

S entiment analysis is an application of Computational Linguistics and Text Mining, in which the hidden emotions in a given text are extracted. In this paper, S entiment Analysis of Malayalam movie reviews is done by classifying the review obtained from user as positive, negative and neutral. A hyid approach for S entiment Analysis is proposed in this work in which Maximum Entropy Model is used for tagging and certain rules are also incorporated to handle certain special cases. ?e Maximum Entropy classifier is a probabilistic classifier which belongs to the class of exponential models. It is based on the Principle of Maximum Entropy. It selects the one which has the largest entropy and from all the models that fit our training data. Maximum Entropy Classification finds out in which class the review must belong, given a context so that it maximizes the entropy of the classification system. ?e rules included ensure that special cases are handled which include negation, intensifiers, dilators etc. ?e system performed well gi ving a considerable precision rate.