Comparative Analysis of KNN, Naïve Bayes and SVM Algorithms for Movie Genres Classification Based on Synopsis.

Abstract

Text classification is a process of categorizing a text into the correct label. Text classification in natural language processing is a challenging task that requires accuracy to get the correct results, manual text classification tends to be inefficient because it requires a lot of time and also experts. The utilization of machine learning for automatic text classification can be a solution to this problem. KNN, Naive Bayes, and SVM are known as some of the most algorithms to solve classification problems, especially text classification. In this study, we are trying to compare the KNN, Naive Bayes, and SVM algorithms for text classification with the problem of classifying movie genres based on a synopsis using datasets obtained from Kaggle.com and IMDB Dataset. The results of this study indicate that of the 12 experiments, Support Vector Machine (SVM) is the bestperforming algorithm with an accuracy of 90%, 93%, 65%, and 63%. It is hoped that this research can help to determine the best algorithm in the text classification process.