Comparative Study of Machine Learning Algorithms for Movie Sentiment Analysis

Movie sentiment analysis is a technique used to analysis sensations, emotions, sentiments, opinion related to the movies. It gained an increasing interest because of the continuous growth of the movies' market. The viewers express their sentiments using comments. The huge quantity of the generated comments requires using machine learning algorithms to ensure the automation of the sentiment analysis to help people to decide which movie worth his time. The collected movies feedbacks are classified into positive and negative In the literature, many solutions have been proposed to apply machine learning algorithms with movie sentiment analysis. Compared to those works, we propose in this article, comparing 35 different algorithms. They are evaluated using the following metrics: Kappa, Accuracy, F1-Score, Recall and Precision. We used two well-known datasets: Cornell Movie Review dataset and Large Movie Review Dataset that are applied in the most of the existing works as benchmarks.