KNN And Naïve Bayes Algorithms for Improving Prediction of Indonesian Film Ratings using Feature Selection Techniques

This electronic document is a "live" template and already defines the components of your paper [title, text, heads, etc.] in its style sheet. *CRITICAL: Do Not Use Symbols, Special Characters, Footnotes, or Math in Paper Title or Abstract. (Abstract) A survey conducted by IDNTimes with 411 respondents shows that audiences in Indonesia choose which films to watch based on several factors. People chose films to watch in theatres based on the genre by 16.4%, actors or actors by 14.5%, and directors by 12.7%. There needs to be more interest from researchers regarding Indonesian film ratings. Most of them discussed foreign films more than Indonesian films. So research is required that discusses Indonesian film ratings. This research will use data mining tools, namely RapidMiner, with the CRISP-DM framework method. The study compared two algorithms, Naïve Bayes and KNN, to predict Indonesian film ratings using the feature selection technique. This research yielded that the KNN algorithm performs better than Naïve Bayes with or without feature selection techniques. Of the three models made for KNN, namely with Forward Selection, Backward Elimination, and without feature selection, all produce the same accuracy value, namely 86.65%. At the same time, Naïve Bayes gets an accuracy value of 85.13% with forward selection, 84.91% with backward elimination, and 71.32% without feature selection technique.