Financial Risk Prediction Model of Listed Companies Based on Particle Swarm Optimization Classifier

When the sample data is unbalanced, the prediction of traditional risk analysis model has tendentious defects, so this paper analyzes the financial crisis influencing factors, and initially establishes the financial early warning index system of China's manufacturing listed companies. Then, particle swarm optimization algorithm is used to improve the parameters of support vector machine to select the optimal index set for financial crisis warning. We also combine the theory and algorithm to perform empirical research and to determine the financial early warning indicators of Listed Companies in China, and use the parameters optimized by SVM and PSO to establish an effective warning model. Compare to the prediction results with the test sample data, the rationality of PSO-SVM model is proved.