Movies Reviews Sentiment Analysis and Classification

Abstract

As humans' opinions help enhance products efficiency, and since the success or the failure of a movie depends on its reviews, there is an increase in the demand and need to build a good sentiment analysis model that classifies movies reviews. In this research, tokenization is employed to transfer the input string into a word vector, stemming is utilized to extract the root of the words, feature selection is conducted to extract the essential words, and finally classification is performed to label reviews as being either positive or negative. A model that makes use of all of the previously mentioned methods is presented. The model is evaluated and compared on eight different classifiers. The model is evaluated on a real-world dataset. In order to compare the eight different classifiers, five different evaluation metrics are utilized. The results show that Random Forest outperforms the other classifiers. Furthermore, Ripper Rule Learning performed the worst on the dataset according to the results attained from the evaluation metrics.