电力企业客户信用风险管理实证研究

随着电力市场改革的不断深入,电力企业客户信用风险管理越来越受到重视。文中介绍了电力企业信用风险管理系统的设计原理及实现方法。重点介绍了如何运用Logistic回归模型对客户进行信用评级以及通过设置过滤器来检测用电异常的方法。目前,该系统已在滁州电力公司得到了应用。应用结果证明,它有助于提高电力企业的信用风险管理水平。

Empirical Research on Credit Risk Management for Electric Power Clients

In order to solve the serious problems of electric charge defaulting and electricity stealing, we design the electric power clients' credit risk management system. The system includes four function subsystems, which are the information management subsystem, the credit scoring subsystem, the credit risk forecasting subsystem and the electro-abnormity detecting subsystem. This paper introduces the design principle and realization techniques of the system, and lays emphasis on how to apply Logistic regression model to credit rating and the approach on detecting electro-abnormity through the filters. In addition, it selects the electricity data of Chuzhou to do the experiment research. The research results show that it can classify the clients correctly, and improve the efficiency of preventing electricity stealing. At present, the system has been applied in electric power company of Chuzhou, and becomes an important credit risk management tool in this company.