摘要
为了提高故障抢修效率和客户服务质量,文中提出了一种基于用户画像技术和故障诊断技术的计量装置故障抢修主动服务预警模型。首先,建立电力用户标签库体系,利用K-Means聚类方法构建用户画像模型,实现用户群体类型的定义;然后,利用极端梯度提升XGBoost算法构建计量装置故障识别模型,实现计量装置故障风险等级的定义。最后,根据用户画像模型和故障识别模型结果,建立计量装置故障抢修主动服务预警模型。通过实证分析,发现故障抢修主动服务预警模型的一级预警等级核查精准率为84.62%,说明该模型具有可靠性。
In order to improve the efficiency of fault repair and customer service quality,an active service early warning model is proposed for metering devices fault repair based on customer portrait technology and fault diagnosis technology.Firstly,a power user tag library system is established,K-Means clustering method is used to construct a customer portrait model,and the definition of user group types is realized.The XGBoost algorithm is used to construct a measurement device fault identification model to achieve the measurement device failure risk level definition.Finally,based on the results of the customer portrait model and the fault identification model,the active service early warning model for the metering device failure repair is established.Through empirical analysis,it is found that the verification accuracy rate of the first-level warning level in the active service early warning model for failure repair was 84.62%,indicating that the model is reliable.
作者
殷新博
唐旭东
王数
陆芸
周建玲
YIN Xin-bo;TANG Xu-dong;WANG Shu;LU Yun;ZHOU Jian-ling(State Grid Changzhou Electric Power Company,Changzhou 213003,Jiangsu Province,China;NARI Group Corporation,Nanjing 210000,China)
出处
《信息技术》
2020年第8期146-151,156,共7页
Information Technology
