点击次数：

论文类型： 会议论文

发表时间： 2017-07-23

发表刊物： 2017 International Conference on Business and Information Management, ICBIM 2017

收录刊物： EI、Scopus

卷号： Part F131932

页面范围： 6-11

摘要： This paper establishes a weight optimization model in credit evaluation, based on the idea that the weight is the most optimal if the credit evaluation result after empowerment has the maximum discriminating power to distinguish default and non-default customers, by establishing nonlinear programming with the maximum discriminating power of the credit scores as the objective function and the sum of weights 1 as constraint conditions. And the optimal weight model is proved by using 1,231 small businesses samples from 28 cities of China. The empirical results show that the discriminating power of credit evaluation result after empowerment by the optimal weight model is the strongest when compared with three kinds of weight models, t value, variation coefficient and mean square error. That means the optional weight model ensure the credit evaluation after empowerment has the maximum discriminating power, and change the disadvantage of existing research do not ensure the evaluation result with the maximum discriminating power after empowerment. ? 2017 Association for Computing Machinery.