Compare Influential Factors Of Chinese And American Box Office And Make Predictions Based On Symbolic Regression

The film industry is one of the most rapidly developing industries in the world.It is not only a means of leisure and entertainment,but also an important way of global cultural exchanges.It is also an important symbol of the economic strength of countries.Although starting later,China’s film market maintains a steady high-speed growth under the new national economic situation and its scale continues to expand.In 2017,the number of screens exceeded 50,000,and the box office exceeded 50 billion.Film is an experiential product with a short product cycle,and the rapidly changing taste of consumers leads to a very uncertain demand for the film.These characters determine that the investment in the film industry is highly risky,but the inflow of capital and resources continues expanding,and production and marketing costs continue increasing.Therefore,it is of great significance to study the influencing factors of box office and the box office forecast.Domestic and foreign scholars have made progress on the box office-related research,but the white box methods,such as queuing theory and multivariate linear regression,are often based on certain assumptions,affecting the reliability of the conclusions;black box methods,such as machine learning,do not build a specific forecasting model which is hard to understand,and some of the research is converted to classification problem,not producing accurate box office values.This article has drawn on advanced domestic and foreign research experience and conducted empirical research based on box office data released in the US and China markets between 2010-2016.First,build a box office influencing factor system including the production budget,distributor,genre,release date,MPAA rating,number of user reviews before release and during the first week of release,the number of theaters or screens engaged in the first week of release,and the box office revenue of the first week,and study the effect of these variables;then,adopt a new method-Symbolic Regression to build models of box office forecast without setting any assumptions,and automatically selecting the explanatory variables while building models before movie release and after the release of the film for one week to forecast the box office respectively.The commonly used method-multivariate linear regression was used as a bench method,it was found that the models constructed by the symbolic regression has better forecast performance;at last,based on national conditions,we compare the influencing factors and forecasting models of the film market between China and the United States.Through research and analysis,it has been found that increasing the number of theaters arranged in the first releasing week of a movie can increase the box office.For movie made in American,the higher the production budget is,the higher the box office is;distributors with better resources and capabilities can improve the revenue for movies made in China;the actors or directors have no significant effect on the box office,and the movie itself is the most important.Based on the above conclusions,this paper provides suggestions for the film industry.