Research on Financial Risk Early Warning in the Online Game Industry Based on Random Forest and Fisher Discriminant Method

Abstract

With the rapid rise of the online game industry and the intensification of market competition, online game enterprises face different degrees of financial risks. It is necessary to establish an effective financial risk early warning model. Based on the online game industry, this paper constructs an index system from six dimensions, and makes an empirical study on financial risk early warning. First of all, a financial risk early warning model based on the random forest is constructed, and the prediction accuracy is as high as 100%, but the machine learning model has poor explanatory ability. Therefore, according to the index of random forest screening, a financial risk warning model based on the Fisher discriminant method is established. This model optimized the prediction effect of the original fisher's discriminant model and the prediction accuracy reached 87.5%. Finally, suggestions are made from the three aspects of cash flow, solvency and profitability of online game enterprises, and national support policies.