Enterprise Financial Risk Early Warning Prediction using Tuna Swarm Optimization Algorithm with Xg boost Model

The number of firms in India going through financial problems has increased significantly in recent times, and the majority of these enterprises lack effective risk management systems. Therefore, it is necessary to set up an early warning system in order to recognize the warning signs of a corporate financial system. Hence, in this research Tuna Swarm Optimization Algorithm with XGBoost (TSOA-XGBoost) is proposed to improve accuracy in corporate enterprise financial risk early warning. In order to verify that the sample data is dimensionally consistent, data samples are initially standardized. Correlation analysis of variables is used to look at both the autocorrelation and the correlation between the variables. The industry's financial issues are then represented by the early warning signal with more screening. Furthermore, intrinsic credit rating is combined with financial data to further boost effectiveness. To enhance the accuracy of XGBoost predictions by decreasing errors, TSO efficiently searches for the best possible parameter combinations. This enables it to provide rapid, reliable risk evaluations to enterprises. According to the results, it illustrated that with an accuracy of 99.96%, the suggested UUBCFA approach performed better than the existing models namely Back Propagation Neural Network (BPNN), Deep Neural Network (DNN), and Graph Regularization–Cox (GR-Cox).