Box Office Forecast Of Domestic Movies Based On XGBoost-CatBoost Combined Model

Film is an important part of China’s cultural industry.It not only entertains the public,but also enriches our spiritual world.It also brings huge benefits to the country’s economic development.China’s film industry is also shifting from high-speed development to highquality development.However,film related parties cannot use scientific methods to effectively estimate the final performance of the film box office,and cannot reasonably plan and adjust the operation management strategy according to the gap between the target box office and the actual box office,which often leads to waste of resources and profit risks.increase.In recent years,with the rapid development of the film industry,film investment has become a favorite project of many investors,but at present,my country’s film industry is still in the developing stage,and only a small part of film investment is profitable.Predicting movies is undoubtedly of great practical significance for risk control.At the same time,by establishing a model to predict the final box office of the movie,and then combining the results of the model to rank the importance,it is of great significance for the producer to improve the quality of the movie and the distributor to rationally arrange the marketing strategy..This paper selects 473 domestic films released from 2015 to 2021 as research samples.According to the principles of data availability and scientificity,from the perspectives of film-related factors,market-related factors and consumer-related factors,selected 22 box office influencing factors,such as the director’s average box office value in recent years,the weighted score of starring works,release schedule,film type,technical effect,film duration,total number of screens,box office on the premiere day,and online search volume are used as predictive indicators.,Maoyan.com,and Baidu Index four websites crawl relevant data and integrate them.Based on these data and referring to the way of previous treatment of categorical variables,all categorical variables are quantified to obtain the training set and test set of the final model.The data set is trained through the two algorithms of XGBoost and CatBoost in machine learning,and the effect is evaluated in terms of prediction accuracy and accuracy.Three methods are used for model combination,among which the least weighted method with relative error has the best effect.According to the model selected in the article,the importance ranking of the factors affecting the box office of domestic films is finally obtained.The results show that the box office of the premiere day,the average box office value of the director in recent years,the Baidu index,the number of people who want to watch,the weighted score of the actor’s work,the cognitive index,the screen The nine indicators of total number,comedy,and Lunar New Year stalls are important factors influencing the total box office of a movie.Among them,in terms of the film itself,the influence of the director and starring actors,the Lunar New Year in the release schedule,and the comedy in the film type are the more important factors.Among the market-related factors,the total number of screens,the box office on the premiere day and the impact of cognitive index The online search volume,ratings and the number of people who want to see are all important among consumer-related factors.In general,in addition to the factors of the film itself,marketing and publicity efforts are more prominent in the impact on the box office.Finally,in the conclusion of this paper,feasible suggestions for film production and distribution promotion are provided.