Comparative Analysis of Models for Movie Review Sentiment Analysis

Movie evaluations represent an individuals' personal viewpoints and possess the potential to influence the prosperity of a film, therefore, precise categorization of such reviews, through sentiment analysis becomes paramount. This paper compares the effectiveness of five different machine learning models for movie review sentiment analysis using two text preprocessing techniques: TF-IDF and Bag-of-Words. The results indicate that Logistic Regression achieves the highest accuracy in both techniques, with Random Forest as the next best-performing model, followed by Gradient Boosting. This study emphasizes the crucial role of selecting the right preprocessing technique and algorithm for a particular task. It also suggests further research into other vectorization techniques and domain-specific features to improve sentiment analysis model accuracy.