基于非监督算法的电力用户催收管理系统设计  

Design of Power User Collection Management System Based on Unsupervised Algorithms

作　　者：段然 胡华俊 黄龙 李培 DUAN Ran;HU Huajun;HUANG Long;LI Pei(Jiangmen Power Supply Bureau of Guangdong Power Grid Co.Ltd.,Jiangmen 529000,China)

机构地区：[1]广东电网有限责任公司江门供电局,广东江门529000

出　　处：《微型电脑应用》2022年第7期168-171,共4页Microcomputer Applications

摘　　要：针对电力营销中欠费回收问题,结合当前的计算机开发技术,提出一种基于非监督算法的催收管理系统。系统设计考虑了以下几方面:一是采用B/S架构对系统进行整体搭建,并重点对系统整体功能框架进行设计;二是引入风险识别模型,以客户历史缴费作为标签,对客户欠费的恶意程度进行识别,最后根据阈值判断该用户是否属于恶意欠费用户;三是针对恶意欠费的用户,采用语音自助播报的方式对用户电费进行催收。最后给出部分实例和恶意识别界面。结果表明,借助该算法,设置不同阈值可得到不同的恶意欠费用户,从而为当前电力用户自助催收提供了一种新的借鉴方式。In view of the problem of arrears collection in current power marketing,a collection management system based on unsupervised algorithm is proposed by combining current computer development technologies.In order to realize the system,this paper mainly proceeds from the following aspects:firstly,the whole system is built by B/S and the overall functional framework of the system is designed;secondly,the risk identification model is introduced to identify the malicious degree of customer arrears with the historical payment of customers as the label,and finally,the threshold is used to judge whether the user belongs to malicious arrears.The third is to collect users’electricity charges by voice self-service broadcasting for malicious arrears.Finally,some examples and malicious identification interface are given.The results show that with the help of this algorithm,different malicious owers can be obtained under different thresholds,which provides a new reference for the current self-service collection of power users.

关 键 词：非监督算法 电费催收 恶意欠费识别 语音催收 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]