Predicting movie success with machine learning techniques: ways to improve accuracy

Abstract

Previous studies on predicting the box-office performance of a movie using machine learning techniques have shown practical levels of predictive accuracy. Their works are technically-and methodologically-oriented, investigating what algorithms are better at predicting the movie performance. However, the accuracy of prediction model can also be elevated by taking other perspectives. For example, it is possible to increase the model accuracy by introducing unexplored features that might be related to the prediction of the outcomes. In this paper, we examine multiple approaches to improve the performance of the prediction model. 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 interpretability of a prediction model. Second, we use an ensemble approach, which has rarely been adopted in the research on predicting box-office performance. As a result, the proposed 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.2

Improving productivity in Hollywood with data science: Using emotional arcs of movies to drive product and service innovation in entertainment industries

“…Ru, Li, Liu, and Chai (2018) explore how incremental daily box office predictions for movies can be generated using deep neural networks analysis. Lee, Park, Kim, and Choi (2018) look at the granular analytics of the movie success using machine learning techniques which is aimed at increasing the accuracy of revenue prediction. Mak and Choo (2018) forecast movie demand using total and split exponential smoothing.…”

How Network Topologies Impact Project Alliance Performance: Evidence from the Movie Industry

“…Movie projects are becoming more complicated, consuming huge financial investments and time; all predictable and unpredictable factors contribute to the high risks and uncertainties. One thread of the literature focuses on the performance of a movie project, which is normally proxied by box office revenue [ 27 , 28 , 29 , 30 , 31 ]. In this line of research, the factors of the fundamental movie profile are considered, such as sequels [ 18 , 25 ], stars [ 17 , 32 , 33 , 34 ], genres [ 35 ], critics [ 15 , 19 ], awards [ 20 , 36 ], culture [ 20 , 37 ], seasonality [ 38 ], and word of mouth [ 16 , 22 , 39 , 40 ].…”

A Machine Learning Approach to Predict Movie Revenue Based on Pre-Released Movie Metadata

“…Making sure that the revenue generated by a film reaches above the cost of making the movie has always been a prime concern for its investors. For example, A Korean film named "Mr. Go" (2013) was estimated to generate a revenue of around 20 million dollars by reaching out to at least 5 million users but ended up reaching only 1.5 million resulting in huge disappointments for the investors (Lee et al, 2018). A system that estimates the box office return of a movie can be a useful tool for the stakeholders in making informed financial decisions and adjusting the marketing strategy to increase the probability of success.…”

“…Also, some studies tried to illustrate the importance of other factors such as violence and horror, in determining the fate of a movie (Gunter, 2018). Also, an attempt has been made to analyze the effects of features such as sequels, number of initial screens, comments regarding a film, presence of stars in films to determine the ultimate fate of the movie (Lee et al, 2018).…”