摘要
随着全采集、全覆盖工作的深入开展,电能表档案信息和运行信息的不断完善,SG186系统的深化应用,全方面珍贵数据的汇总为电能表资产寿命预测提供了条件。电能表故障概率的预测对工作计划的制定有着指导性意义,关系到资产到货、检定、配送、仓储等多个环节。本文以CART算法为基础,通过分析与电能表故障相关的多个属性,建立并优化决策树,给出电能表故障概率,为继续开展此项工作提供了理论基础和数据模型。
All aspects of the great summary, of valuable data provides possibility for the prediction of power meter asset life with the indepth development of full collection and full coverage work, the continuous improvement of power meter file information and running information and further application of SG186 system. The prediction of power meter fault probability has guiding significance for the formulation of work plan, which is related to assets arrival, verification, distribution, warehousing and so on. Based on the CART algorithm, this paper establishes and optimizes the decision tree by analyzing multiple attributes related to power meter fault. The fault probability is given, which provides the theoretical basis and data model for continuation of this work.
出处
《电力大数据》
2017年第10期7-10,60,共5页
Power Systems and Big Data
关键词
电能表
决策树
故障概率
power mete
decision tree
fault probability
