| Since the 1990s, along with the development of Chinese economy and continuous expansion of domestic demand , individual consumption credit has been developing rapidly in China. In order to prevent credit risks more efficiently and promote the development of individual consumption credit business furthermore, commercial bank needs to establish an effective, scientific and reasonable individual credit evaluation architecture.As one of the most widely used classification algorithms of data mining, decision tree algorithm is characterized by high accuracy, simplicity, and high efficiency. So it is very suitable for the application in the individual credit evaluation system of commercial bank.In this paper, we acquire the individual credit data of a commercial bank originated from the internet, use the decision tree algorithm based on C5.0, through data pre-processing, selection of training sample and then establish the individual credit evaluation -decision tree model. According to the actual demands of commercial banks, we also adjust the cost-matrix and pruning degree of our model which finally leads to a modified model. From the classification and forecasting result of our model, we can get the conclusion that the individual credit evaluation model based on decision tree method has the advantage of high accuracy, low risk cost and high maneuverability.Meanwhile, we select two other representative methods—logistic regression and neural network to build individual credit evaluation models. The classification and forecasting results compared with previous decision tree model's shows: the logistic regression model is the most stable; the neural network model is the most accurate; while the decision tree model has the lowest risk cost, is the most efficient and presents the best comprehensive performance.At the end of the paper, we also bring some suggestions on the establishment of individual credit evaluation architecture for commercial banks. |
PDF Full Text Request