"Using decision trees to characterize and predict movie profitability on the US market".

Abstract

The filmmaking is one of the most important branches of the entertainment industry primarily because of the huge revenues that it generates. The producer plays an essential role in filmmaking, as they provide the funding required to turn out quality blockbusters for cinemagoers. Film production is a risky business, as illustrated by the examples of films that fail to cover costs every year. In this respect, tools capable of predicting movie profitability are of potential use to producers as a decision-making tool for deciding whether or not to produce a movie project. In this paper we report a study using historical data on over 100 films produced in the United States (including their genre, opening month, duration, budget, etc.). Decision trees were extracted from these data in order to forecast whether or not a film will be profitable even before it is produced. Decision trees are models commonly used in the field of artificial intelligence as decision support tools.The results show that the resulting model forecasts whether or not a movie will be profitable with an accuracy of over 70%, and this model can be used as a decision support tool for film producers. The proposed approach is not designed to be used as a standalone tool; it should rather round out other forecasting methods, including producers’ foresight and judgement. The approach presented here could be equally applicable to other branches of the entertainment business, such as the music or video game industries.