Comparison of support vector machine and support vector regression: An application to predict financial distress and bankruptcy

Lately, many notorious financial distress and bankruptcy events occurred in the world economic. As we known, bankruptcy of Lehman Brothers Holdings Inc. (LEH) is the largest bankruptcy filing in U.S. history in 2008. These events have serious impacted on the socio-economic and investment in public wealth. Due to solve this dilemma, this research collected 68 listed companies as the raw data from Taiwan Stock Exchange Corporation (TSEC). The support vector machine (SVM) and support vector regression (SVR) techniques were used to implement the financial distress prediction model. Moreover, we adopted a total of 22 ratios which composed of 13 financial ratios and 9 macroeconomic indexes to be the input variables for these models. Finally, the experiments obtained the accuracy rate, Type II error rate and RMSE (root mean squared error) of these classification methods for the financial distress and bankruptcy prediction.