Using Machine Learning to Identify Movie Genres through Online Movie Synopses

Abstract

Movie synopsis is important for audience to know about a movie within a short time. A good movie synopsis should reflect the genre, structure and main plot of a certain movie. The aim of this paper is to use machine learning to identify the genres of movie through movie synopses. The movies and corresponding synopses in database are downloaded from the Kaggle and ROTTEN TOMATOES websites. This study uses two supervised learning models (k-NN and SVM) and two deep learning models (CNN and RNN) to classify the genres of movie through movie synopses. Secondly, it tries to eliminate the interference by actively eliminating proper nouns. Finally, it compares and analyzes the performance of all models in different training sets. The result is that RNN with LSTM layer is the most suitable model for analyzing a large number of text for movie synopses, and the accuracy of judging movie genres is 80.5%. This study promotes the understanding of machine learning model selection in the adoption of movie genres classification based on movie synopses.