Enhanced Financial Distress Prediction Model Using GWO Adaboost Optimized Support Vector Machine

Abstract

With the continuous changes in the global economy and increasingly fierce market competition, financial risks faced by municipal companies are becoming increasingly complex and diverse. To provide more accurate and timely risk warning for listed companies and support risk management, a financial distress prediction model is constructed using support vector machine (SVM) in machine learning algorithms to forecast financial data risks. Grey wolf optimization (GWO) algorithm is combined with adaptive boosting (Adaboost) algorithm to optimize parameters. The study is based on the CSMAR database of Guotai An and the data of listed companies on the official website of the National Bureau of Statistics from 2019 to 2024 as samples, and selects normal and special treatment (ST) company data in the A-share market for comprehensive analysis. The listed companies in the dataset include normal companies and ST companies, with a ratio of 1:1. The findings indicate that the prediction accuracy, recall rate, and F1 value of the research model reach 92.67%, 93.52%, and 93.09%, respectively. In the prediction of normal enterprises and ST enterprises, the prediction errors of the improved SVM model are 2.67% and 3.22%, respectively. In summary, research on financial distress prediction and risk management of listed companies based on machine learning can provide more accurate and timely risk warnings for enterprises, help identify potential financial distress in advance, and provide strong support for enterprise decision-making.