A credit scoring model using Support Vector Machine

A novel credit scoring model using the Support Vectors Machine (SVM) is proposed. The credits of 106 listed companies of China in 2000 and further forecast 13 pre-lost companies in 2001 are evaluated with this model. The 106 listed companies are divided into two groups: the "good" group and the "bad" group, according to their performance. Four primary financial indexes, which are income per share, net asset per share, return rate of net asset, and cash flow per share, are considered for each listed company. The simulation results show that a high classification correct rate of up to 98.11% is attained with the SVM credit scoring model. Moreover, it possesses strong adaptive ability, so it can be used to predict financial distress of the companies. The forecasting results show that the forecasting accuracy rate of the model reaches 100%.