Compared Study on Enterprise Bankruptcy Forecasting Method

摘要

Enterprise bankruptcy forecasting is one of most significant content to manage credit risk,which can be solved using classification method. By now, there has three typical classification methods to forecast enterprise bankruptcy: the statistics method, the Artificial Neural Network method and the kernel-based learning method. The paper introduces the first two methods briefly, then introduces the theory of Support Vector Machine (SVM), lastly build the corresponding models to compare the precision of three methods to forecast enterprise bankruptcy with the China Stock Exchange data. From the positive analysis we can draw a conclusion that the SVM method has a more adaptability and precision to forecast enterprise bankruptcy.