Rate movie app: Implementation of k-nearest neighbors algorithm in the development of decision support system for Philippine movie rating and classification

Abstract

Movies that are publicly exhibited in the Philippine Cinema, regardless if produced locally (local films) and/or outside the country (foreign films) undergo a thorough evaluation before public exhibition to properly identify suited audiences. There are many factors that contribute to the classification and rating of a specific movie. Movies play a vital role for Filipino culture as for some people; these serve as their leisure activity, for other people, these are not just a leisure activity instead a form of visual art that may send important messages to the audiences and/or may re-enact human personal experiences. It is very important that movie(s) will be classified accordingly without any form of biases. This paper promotes a Decision Support System that can be used in predicting movie classification and rating using historically evaluated movies from 2010 to 2017. The study considers the user ratings on the following attributes: Sex & Nudity, Violence & Gore, Profanity, Alcohol, Drugs & Smoking and Frightening and Intense Scenes scrapped from a public movie database. Along with these considerations are the genre(s) associated with a movie. The study conducted revealed that K-Nearest Neighbors Algorithm outperforms Naive Bayes and J48/C4.5 Algorithm in classifying Philippine Movie rating with 92.80% accuracy as compared to 68.70% and 56.79% for Naive Bayes and J48/C4.5 algorithm respectively. The developed decision support system implements the K-Nearest Neighbors algorithm to satisfy the objectives mentioned. With this, Review Committees who evaluate movies may have guides in making critical decisions in the domain of movie evaluation.