Implementation of Sentiment Analysis Movie Review based on IMDB with Naive Bayes Using Information Gain on Feature Selection
Meta Mahyarani, Adiwijaya Adiwijaya, Said Al Faraby, Mahendra Dwifebri · 2021
A movie review is an information that forms an evaluation relating to each aspect inherent in a movie. The information contained in the film can be concluded as the quality of a film experience from the audience, but the rating of 'inappropriate' with the context sentence makes the rating of the film 'not recommended'. This issue supports this research where a film should be classified by sentiment. This paper uses the Naïve Bayes method for classification, TF-IDF as the feature extraction, and Information Gain as the feature selection. However, the uses of Naïve Bayes to classify the word are conditionally independent of each other. This research uses TF-IDF and Information Gain to prevent this problem. From the evaluation results on the testing process, the maximum performance results are 84.50% for precision, 88.27% for precision, 88.27% for recall, and 86.34% for f1-score. From the evaluation results, it can be concluded that using TF-IDF, Information Gain, and Naïve Bayes can produce precision, recall, and f1-score value which is good enough.