Design of Financial Risk Prediction System for Listed Enterprises Based on Improved Grey Wolf Algorithm and BP Neural Network

The existing financial crisis early warning indicator system relies too much on financial indicators and ignores non-financial indicators. Non-financial indicators, such as corporate governance, are affected by the difficulty of information acquisition, data accuracy, subjective judgment and other factors, which makes it difficult for the early warning system to accurately capture and assess the real situation of the company, and then generate bias and fail to comprehensively reflect the real situation of the enterprise. Therefore, we propose a back propagation (BP) neural network prediction method for financial crisis risk of listed enterprises considering corporate governance. Corporate governance indicators are set from the aspects of board governance, management incentives, and equity structure arrangement. Together with the financial indicators as the prediction indicators of the financial crisis risk of listed enterprises, we use the principal component analysis (PCA) method to filter out the indicators that have an important influence on the prediction. Constructing a risk prediction model based on the improved Grey Wolf algorithm and BP neural network, and optimizing the setting of the Grey Wolf algorithm to achieve the minimum mean square of risk prediction error and BP neural network parameters, effectively overcoming the problems of traditional BP neural networks being prone to local optima and slow convergence speed. Based on the relationship between predictive indicators and financial crisis risk status, predict the financial crisis risk status of listed companies. The experimental results show that the method is accurate in predicting the financial crisis risk of different types of listed enterprises, and the prediction results match the actual risk status.