MPAA Rating Prediction Using Script Analysis for Movies Using Ensemble Modeling

Abstract

Movies serve as a form of artistic expression, utilizing the medium of film to communicate ideas, stories, emotions, and atmosphere to create a specific experience for the viewer. The messages and themes in movies have the power to shape social attitudes and potentially bring about change. However, it is important to consider the potential negative effects that certain scenes and dialogues can have on vulnerable members of society, such as children. They can be considered a reflection of society’s current attitudes. The Motion Picture Association film rating system (MPAA) has been utilized since 1945 to classify movies and TV series based on factors such as violence, language, and sexual content. This research aims to use data analysis and machine learning to predict the MPAA rating of a movie based on its script by splitting the script into categories such as “angry,” “sad,” “happy,” “surprised,” and “fear” and examining the impact of each on the rating. An ensemble method is proposed to improve the accuracy of the prediction. Additionally, a bidirectional LSTM-based model with attention is proposed to classify the movie into one of five classes: R, G, PG, PG-13, and NC-17.