Predicting Movie Success with Machine Learning Techniques: Theoretical and Methodological Approaches to Improve Model Performance

Abstract

Previous studies on predicting the box-office performance of a movie using machine learning techniques have shown practical levels of predictive accuracy. However, their efforts to improve the model accuracy have been limited only to the methodological perspective. In this paper, we combine a theory-driven approach and a methodology-driven approach to further increase the accuracy of prediction models. First, we add a new feature derived from the theory of transmedia storytelling. Such theory-driven feature selection not only increases the forecast accuracy, but also enhances the explanatory power of the prediction model. Second, we use an ensemble approach, which has rarely been adopted in the research on predicting box-office performance. As a result, our model, Cinema Ensemble Model (CEM), outperforms the prediction models from the past studies using machine learning algorithms. We suggest that CEM can be extensively used for industrial experts as a powerful tool for improving decision-making process.