Sentiment Analysis of Movie Reviews Using Support Vector Machine Classifier with Linear Kernel Function

Abstract

Sentiment analysis refers to the process of determining the opinion stated by the user corresponds to positive, and to be negative or considered to be neutral. The mechanism of sentiment analysis is said to be the process of opinion mining which in then resembles the behavior/attitude measurement of the speaker. This is extremely useful in a place which there is a complete need for a recommendation for the user to follow a specific case of action. In public domain, the aspect of sentiment analysis is helpful for the user to state a specific nature of the action. This research work focuses on the analysis of review data to determine the aspect based on sentiments using TF, IDF, and SVM. The model extracts the textual reviews and classifiers them into positive, negative, and neutral cases. The result retrieved with the proposed scheme gives an improved accuracy of about 87.56% determining the positive and negative cases more efficiently. With this proposed approach the classification of review data can be made more efficiently for various sort of recommendation systems which makes the user have good insight for a product review, movie review, and user rating analysis.