Методика нейросетевого прогнозирования кассовых сборов кинофильмов
Theme. Neural network forecasting in the film business. Goal. The article is devoted to application of economic-mathematical modeling in the field of film industry, in particular – to predict revenue and profit from distribution of future films, the identification of factors influencing the commercial success of the film business. Methodology. The basis of economic-mathematical model is a neural network trained on known historical data about the rental of movies, including 20 of the input parameters describing the production costs of the film, its duration, the characteristics of the Director, actors, characteristics of the plot of the film, country-made, genre, etc. Results. RMS relative error of the model was 13.8%, the coefficient of determination of 0.86. Model capabilities are demonstrated in the films: "the Da Vinci Code", "Star wars." Computer experiments were performed by the method of "freezing": using neural networks was carried out computations for the virtual change of the input parameters of the model, for example – the budget of the film, while the other input parameters remained unchanged. It turned out that the virtual increase in the budget of the films has a different effect on the predicted box office films, as well as on the profit margin. In the first case, the virtual increase in the budget leads to a significant increase in projected box office and profit, whereas in the second case, cash collections from a certain point ceases to increase, and profit growth is slowing down and even observed her fall. Different effects on the success of the movie business and other settings of the films. On the basis of computer experiments offered a number of recommendations that could contribute to increased box office receipts of the investigated films. Conclusions. The value of this study is that the created economic-mathematical model can be used for optimization of costs and selection of parameters when planning new movies. It allows you to make forecasts of box office receipts and profits from the rental of newly created films, as well as to investigate the influence of various parameters on the commercial result of the film business.