Machine Learning Based Sentiment Analysis on Movie Reviews

Abstract

Sentiment analysis is a recent area of research where vast amounts of data are analyzed to offer insightful information about a particular subject. It is a strong instrument that can benefit organizations, customers, and even governments. Nowadays, analyzing the emotional content of movie reviews is popular. The approach heavily relies on textual emotion recognition. Analysts in Machine Learning (ML) and Natural Language Processing (NLP) have investigated a variety of approaches to execute the procedure with the highest level of accuracy. There are three stages to the sentiment analysis procedure for movie reviews. We must first gather reviews from online platforms. The collected data will then be analyzed. Finally, we will have all of the data that has been processed regarding the tone of those reviews. The results of the model analysis may aid the consumer in understanding how viewers feel about a certain film. To conduct this research, movie reviews were subjected to sentiment classification methods. For review sentiment classification, we looked at SVM, Naive Bayes, Logistic Regression, and K-NN, four supervised machine learning methods. Empirical results with a large number of reviews in the training dataset show that the SVM model performs better than the Naive Bayes, Logistic Regression, and K-NN approaches. The SVM method achieved an 87.46% accuracy, an 87.21% precision rate, and an 87.46% recall rate.