| In the past decade,China’s cultural industry has shown a rapid development.In2022 alone,the cultural industry and its related revenue reached 12,180.5 billion yuan.Different from other industries,culture has strong regional and dispersed characteristics.Therefore,while considering the analysis and prediction of regional industrial development,the overall analysis and prediction of the overall development of geographically dispersed animation,games(e-sports)and cultural industries can help the government and cultural industry parks to make scientific management decisions on the development of local cultural industries.Therefore,it is of great significance to build an information platform that can analyze industrial data from multiple angles and predict the development trend of animation,games(e-sports)and culture industry.One of the main problems faced in the prediction and analysis of cultural industry is that the geographically dispersed industrial layout makes the cultural industry data have the characteristics of hierarchical time series data from the perspective of administrative division.However,the current analysis and prediction technology does not take the hierarchical time series into consideration,which is mainly reflected in the following aspects: The traditional method is used to predict the hierarchical time series and the traditional mathematical method is used to reconcile the hierarchical results.Therefore,how to make a more intelligent and scientific prediction of hierarchical cultural industry data is one of the main problems to be solved urgently.This thesis takes the animation game(e-sports)industry as the main analysis object,because the above industry is a characteristic industry of Chinese culture.In view of the above characteristic industries,this thesis studies the development status of cultural industries at home and abroad in detail and analyzes the problems existing in the current characteristic industries.In view of these problems,combined with the actual needs of users using the cultural industry analysis and prediction system,a deep learning-based animation game(e-sports)industry analysis and prediction system is designed and implemented.The system includes eight major functional modules: data management,user management,macro development analysis and forecast,regional development analysis and forecast,academic hot spots analysis and forecast,investment trend analysis and forecast,industrial information visualization analysis and cultural industry innovation ability analysis.This system realizes the visualization analysis and development trend prediction of animation game(e-sports)industry.In the development trend prediction function of animation game(e-sports)industry,a hierarchical time series prediction model based on deep learning is proposed in order to solve the problems of low accuracy of single-layer prediction and tedious coordination of inter-layer results in the current hierarchical time series data prediction.The model divides the hierarchical time series prediction problem into two parts to solve: First,the hierarchical time series is decomposed according to the hierarchy.Because time series at different levels have different data characteristics,for example,top-level time series data has higher relative stability,less noise and is easy to model,while bottom-level time series data has more noise and features are difficult to capture.To solve this problem,the adaptive expansion coefficient is proposed in this paper on the basis of the use of time series convolutional neural network,and a functional relationship is established between the information of the level where the time series is located and the expansion coefficient,so as to realize the adaptive change of the network structure for the time series of different levels,and ensure the targeted and accurate prediction of the time series of each level.Secondly,the hierarchical time series has the consistency constraint,while the traditional inter-layer harmonic method mainly relies on the mathematical method,so the process is complicated and the harmonic effect is limited.To solve this problem,this paper constructs an interlayer error harmonic based on the deep neural network,and realizes the harmonic of the results between different levels through the deep neural network,so as to ensure that the final predicted results meet the consistency constraints between levels.The animation game(e-sports)industry analysis and prediction system based on deep learning uses browser/server architecture as a whole,and uses Python language and Django framework for system development.The design mode of background business realization part is MVT mode,and the important information in the whole system is stored through My SQL database.This paper makes detailed functional and performance testing of the developed system.The test results show that the system developed in this paper meets the expected development needs,and can solve the problems in the analysis and prediction of cultural industry to a certain extent,and plays a certain promoting role in our country’s cultural industry development. |
