Empirical Analysis On The Prediction Of Online Game Revenue Based On Machine Learning Algorithms

According to the report on China's game industry in 2018,the actual sales revenue of China's game market in 2018 reached 214.44 billion yuan,accounting for 23.6% of the global total.China has become an important part of the global game market.With the rapid development of the Internet market,the influx of capital and the improvement and support of relevant policies,the game industry has become an important part of China's market economy.As a micro subject under the market economy environment,enterprises aim to pursue the economic benefits of products.Accurately predicting the amount paid by users of a game is to accurately predict its profitability,which is of great significance to the internal resource integration and relevant management decisions of game enterprises.In this paper,intelligent algorithms such as BP neural network,least square support vector machine(LSSVM)and random forest(RF)as the means of a certain game top-up amount for regression prediction.In order to eliminate the influence caused by different dimensions in the original data,the z-score standardization was used to preprocess the data in advance in this paper,and then the principal component analysis method(PCA)was used to conduct dimensional reduction processing on the data to reduce the operation workload.Finally,the linear model is introduced and compared with the three intelligent algorithms.Through the calculation of MAE,RMSE and MAPE,it is found that random forest model performs well in prediction.The accurate prediction of the payment amount of users in the game is of great significance to the integration of enterprise resources and the making of management decisions.