Comparison of feature selection method in movie classification utilizing Naïve Bayes classifier

Abstract

Movies have been currently favored by Indonesians as a means of entertainment medium. Movie fans cover among young adult from all walks of life from young children to the elderly. With the increasing number of fans in Indonesian movies, indirectly encouraging the annual productions of movies. This phenomenon encourages people confusion in watching and preferring about the genre or type of film. Based on this research, it is thus possible to east the problem. This effort assists in grouping or classifying a film. To perform this classification, this study utilizes the Naïve Bayes classification method combining with the selection feature information and the chi square. In this case, the selection of the user features will be compared, which is useful to obtain the best classification value. The results of the classification of films based on synopsis that has the highest results utilize the Naïve Bayes classifier and feature selection of chi square with an accuracy of 90%. The average value of precision is 89%, and the average value of recall is 88%.