Intelligent System for Prediction Box Office of the Film

P. 18–25.

Yasnitsky L., Mitrofanov I., Immis M.

Translator: I. Mitrofanov, M. Immis

Abstract. This article describes a mathematical model designed to predict the
box office of movies. The model is based on a neural network trained on data
about films obtained from open sources. Computer experiments were performed
by the method of “freezing”: with the help of a neural network calculations were
performed with a gradual change in the value of one of the input parameters of
the model, while the remaining input parameters were not changed. It is
established that the size of the film budget has the greatest impact on the amount
of box office among all other input parameters. However, its impact is not
always positive. It is established that the United States, as a country that takes
part in the production of the film, is able to have the greatest impact on box
office compared to other countries. According to the research the duration of the
film to varying degrees can affect the amount of box office. Moreover, the power
of influence depends on such a factor as the genre of the film. As a rule, the
power of this influence increases with the increasing semantic load of the film.
The practical value of the study is that the created mathematical model can be
used to optimize costs when planning the production of new films.

Keywords: box office, genre, profitability, scenario forecasting, Film business, adaptive neural networks

In book

Integrated Science in Digital Age: ICIS 2019. Lecture Notes in Networks and Systems

Т. 78. , Springer, 2020.