Study On Box-office Prediction Of Imported Film Using Film Evaluation

Prediction of movie box-office revenue is an important means which ensures the return on investment and controls the risk of issuing movie. Imported film is the main force of domestic box-office revenue in Chinese film market. As the imported film plays an indispensable role in the game process in film market, it is important to make prediction for its box-office revenue. Before the film companies want to import a movie, the prediction of box office helps companies to decide whether the movie is worth to issue or not. In accordance with the trends of box office of imported film, cinema can adjust the strategies to control costs after the movie released.Being an important influence on film consumption behavior, film evaluation is a reliable source of information that helps audiences choose favorite films, and provides reference for producers and issuers of film. The degree of film evaluation reflects acceptance of audiences to the film and indirectly dominates watching behavior of audiences, which affects box-office revenue to some extent.The prediction technology is highly advanced based on the support of data mining and statistical analysis. However, the researches on the box-office predictions, especially prediction of imported film are fewer. Due to the limitations of the irregular data and low accuracy, the applications of box prediction are even scarcer in actual project.Firstly, this paper described the status of Chinese film market and the significance of box-office prediction. Domestic box-office of imported films in recent years and the effects of film evaluation on box-office was discussed. Secondly, current mainstream prediction research method was introduced. Domestic box-office and film evaluation data from oversea websites, including rating and review data, was collected. The relevant chapter introduced the procedure, such as extracting film feature, preprocessing data, quantifying feature, establishing structure. Thirdly, this thesis built box-office prediction model based on regression analysis and correlation analysis respectively. The feature of box-office and evaluation was added to the models to make comparison between prediction in different situations. Two prediction models showed better prediction accuracy in respective applicable domain. The subject achieves the prediction models in an actual engineering application and research prospect.