Application of supervised machine learning algorithms in diagnosis of abnormal voltage

摘要：

　　This paper introduce a typical supervised machine learning algorithm, which is called C4.5 decision tree classification algorithm. On the basis of analyzing the reason of metering device abnormal voltage, we established a metering device voltage judgment model to evaluate the model performance, then use this model screen out the measurement anomaly, the on-site test results verified the accuracy of the model. This method can effectively improve the accuracy of this kind of abnormal judgment and avoid failure, improve the work efficiency of on-site troubleshooting.

关键词： Intelligent watt-hour meter Electricity information collection system Machine learning decision tree abnormal voltage

作者: Li Yifei Wu He Pang Shuai Song Weiqiong Ding Ning Wang Fang

作者单位: State Grid Beijing Electric Power Research Institute,Beijing 100045,China

会议类型: 国际会议

会议名称: 2016中国国际供电会议(CICED2016)

会议地点: 西安

会议语种:英文

页码: 1-5

在线出版日期: 2016-09-01（万方平台首次上网日期，不代表论文的发表时间）