Research and Verification Analysis on Early Warning Model through Support Vector Machine Algorithm

Abstract

Financial crisis early-warning model has always been the research focus of domestic and foreign scholars. Based on the review of related research at home and abroad, by using the support vector machine (SVM) method, taking the return on equity, turnover of accounts receivable, quick ratio, profitability cash ratio, high and new technology product and service revenue growth rate were used as input variables, and defining the “default” as the output variable, a financial crisis early-warning model of Chinese listed companies was established. Later, using the data of 75 non-listed listed companies for empirical analysis, the accuracy of 98% of training samples and 9% of verification samples were obtained, with a relatively high prediction accuracy, which proved the effectiveness of the SVM method in the financial crisis early-warning modeling of listed companies.