A study of personal credit scoring models based on fuzzy ART
Selecting the right method for personal credit scoring has great significance for risk prevention in personal credit business which is owned by commercial banks. A model for personal credit scoring is built in this paper using Fuzzy ART, 100 training samples and 150 test samples are calculated and analyzed in the model. The results show that erroneous judgement rate and flawing rate of the model are reasonable, the overall judgement accuracy is 88.67%, and the judgement accuracy of default risk can reach 92.00%, which is much higher than the current logistic regression model, linear programming model, BP neural network model.
Author's Name: Jiang, M., Lin, S.
Volume: Volume 6
Issues: Issue 9
Keywords: Credit scoring, Fuzzy ART, Personal credit