Research And Implementation Of Movie Box Office Prediction Model Based On Neural Network

In recent decades,the film industry has developed rapidly,both in terms of the number of films and the cost of film production,showing an increasing trend year by year.The box office prediction has been taken as an urgent concern for the movie industry by the huge movie market scale and high movie production cost.As early as the 80s of the last century,the industry began to study the box office prediction problem.The box office prediction model is constructed by empirical analysis of factors affecting the box office of the movie.The development of big data,predictive models and methods provide data and technical support for box office predictions.From the point of view of box office forecasting indicators,the development of the Internet has also led researchers to focus on mass online content such as word of mouth,social net-working sites,and search volumes of user behaviors,and apply them to box of-fice forecasting.However,the selection of box office forecast indicators needs to consider data availability,quantification methods and quantization accuracy.From the perspective of box office prediction methods,machine learning has become the most popular prediction model.But most machine-based box office forecasting studies convert box office predictions into classification problems,implementing interval estimates rather than point estimates.In view of the above background and existing problems,this thesis stud-ies the movie box office prediction based on neural network.In the design of predictive index system,it combines the basic characteristics of film,word of mouth and search volumes of user behaviors,and defines the quantitative method;In the choice of prediction method,BP neural network is a multilayer feedforward neural network trained by error back propagation algorithm.It is applied to the movie box office prediction problem,and the non-linear mapping relationship between the box office prediction indicators and the box office is explored to realize the prediction.However,the BP neural network is easy to fall into the local minimum.To solve this problem,the genetic algorithm is introduced to optimize the selection of the weights and thresholds of the BP neural network to realize the prediction of the movie box office value.The ex-perimental results show that the model has a good box office prediction effect and the average absolute error rate is less than 15%.Besides,the prediction results are obviously better than unoptimized BP neural network model.