Prediction of Gross Movie Revenue in the Turkish Box Office Using Machine Learning Techniques

Abstract

The gross revenue of a movie in the box office has been a concern of the movie industry. In the last few years, there have been studies on predicting various movie attributes. The field lacks a gross movie revenue prediction model that specifically concerns the gross movie revenues in the Turkish box office. The aim of this study is to build a model to predict the gross movie revenue in the Turkish box office using machine learning techniques. This study is conducted on 150 movies that were in the Turkish box office in 2018. The techniques involved multiple regression analysis including the ridge regression and the lasso, tree-based methods including random forest and boosting, SVM and KNN regression. All models were built using the R programming language. Methods were compared using their MSE values. The lowest MSE was obtained with the Random Forest model.