MOVIE GENRE PREDICTION USING MACHINE LEARNING

Abstract

When creating new films, information often leaks. Such information is incomplete and can reveal details like who is working on the film, when it will premiere, the film's title, etc. To get a complete picture of the film, it's necessary to know the genre it belongs to. In this study, genre prediction is performed using Multilabel classifiers (Multilabel k-Nearest Neighbors, Classifier Chains, Binary Relevance) based on information about the people working on the film, their roles, the film's title, the year of its premiere, and whether the film is intended for children. The prediction results are evaluated using Micro-average and Macro-average evaluation methods.