Sentiment Analysis on Movie Review from Rotten Tomatoes Using Logistic Regression and Information Gain Feature Selection

Abstract

The advancement and development of technology today can have a positive influence on the use of the internet and also on the dissemination of information it contains, including information about the world of cinema. With this convenience, there are many movie reviews that can be obtained easily. Movie reviews are very influential in the various ways movies are available. Thanks to the ease of various information on the internet, the number of movie reviews has become diverse. Therefore, it is necessary to do a sentiment analysis. In this research, the classification method used is Logistic Regression. The method was chosen because it has accurate classification accuracy. In this study, Information Gain was also chosen as a feature selection because it is good enough to do a filter approach in classification. Furthermore, for feature extraction, TF-IDF was chosen because it can overcome data imbalance in the dataset. The best model resulting from this research is a model built without using stemming in the preprocessing stage, without using information gain feature selection, and using parameters in Logistic Regression which produces an f1-score of 76.50%.