基于BP神经网络电力大客户信用等级评价研究  

Large power customer credit rating based on BP neural network

作　　者：王宇哲[1] 雷霞[1] 陈晓盛[1] 黄贵鸿 徐贵阳[1] 

机构地区：[1]西华大学电气与电子信息学院,成都610039

出　　处：《电力需求侧管理》2015年第5期49-53,共5页Power Demand Side Management

摘　　要：电费回收风险是供电企业电费安全风险管理的重要环节,而电力大客户的信用管理是规避电费回收风险的重要组成部分。为了能够更好地为供电企业规避风险,为电力客户提供差异化服务提供有力的依据,从供电企业的角度出发,根据电力客户行业的发展情况、客户的历史信用情况以及客户财务3个方面构建指标体系,并通过层次分析模型将电力大客户分成5个信用等级作为BP神经网络训练的目标值,通过训练得到的BP神经网络,能够对企业的信用等级进行评价,最后通过收集某市电力公司数据验证了该方法的可行性。The risk of electricity charges collection plays an important role in electricity safety risk management for power sup- ply enterprises, while the aversion of the former partially depends on large power customers＇ credit management. In order to help en- terprises reduce risks and provide evidence for offering power cus- tomers diversified services, this paper constructs an index system based on the situation of power customer industry, the historical credit record of customers and enterprises＇ finance. And by mathe- matical method of AHP, this paper classifies large power custom- ers into five credit ratings as the objective value of BP neural net- work. BP neural network can make assessments of enterprises＇ credit ratings. Finally, through data collection in electric power company, it proves the feasibility of the method. Initially sets up the early warning system of large power customer risk management.

关 键 词：电力大客户 信用等级 层次分析法 BP神经网络 风险预警管理系统 

分 类 号：F407.61[经济管理—产业经济]