LS-SVM for bad debt risk assessment in enterprises

With the development of market economy in China, the problem of bad debt becomes increasingly serious in enterprises. In this paper, a bad-debt-risk evaluation model is established based on LS-SVM classifier, using a new set of index system which combines financial factors with non-financial factors on the basis of the 5C system evaluation method. The bad debt rating is separated into four classes- normality, attention, doubt and loss through analyzing accounts payable. Then the LS-SVM classifier is trained with 220 samples which are stochastically extracted from listed companies of China in industry, and the four classes are identified by the trained classifier using 80 samples. Then, BP neural network is also used to assess the same data. The experiment results show that LS-SVM has an excellent performance on training accuracy and reliability in credit risk assessment and achieves better performance than BP neural network.