No abstract

Various sentiment analysis methods have been proposed to obtain reviewing opinions from machine learning and vocabulary-based sentiment. In previous research using only machine learning methods, I got an unsatisfactory accuracy score because a lexical / lexicon hybrid method was used in this study. Machine learning is proposed that can improve the performance of existing sentiment analysis. With this proposal the researcher was able to obtain experimental results that showed that the classifiers that produced the best results for the game review were Hybrid Lexicon and Nave Bayes with an accuracy of 70%.