首页|Multiple Criteria Decision Making based Credit Risk Prediction using Optimal Cut-off Point Approach
Multiple Criteria Decision Making based Credit Risk Prediction using Optimal Cut-off Point Approach
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NSTL
Credit risk assessment is an important task for banks and financial institutions as the loan defaulters and market competition has been increased. In the competitive market, data mining proved to be an efficient technique for predicting credit risks accurately. But in practice, the data mining classifiers use the default cut off value of 0.5 to predict the binary outcomes. If the group size is not equal in datasets, then the cut-off point of 0.5 is not suitable. For imbalance multiple criteria classification problems, it is essential to determine the optimal probability cut-off point. In this paper, a different approach, MCDM (Multiple Criteria Decision Making) based optimal probability cut-off point is proposed for predicting credit risk using hybrid data mining methods. The proposed approach is more advantageous than the usual approach because it determines the optimal cut-off point with improved correct classifications and it also considers MCDM methods for choosing the optimal cut-off point which is very effective in decision making process among multiple criteria.
NSTL
