摘要
针对智能配用电大数据中数据质量差,造成用电异常分析结果不准确的问题,提出基于最小二乘法和聚类的用电数据异常分析算法,首先,对用户用电数据进行分析,并将数据特征进行规范化,其次,通过对已有聚类算法特点分析,采用K-means算法实现数据分析,并利用最小二乘法对数据点进行拟合,计算离心点数据的阈值,并将离心点数据加入噪声集进行隔离,从而提高K-means算法的效率,最后,将传统的K-means聚类算法与该算法进行比较,验证了该算法在准确率和误报率方面都取得了较好效果。
Aiming at the problem that the data quality in the power distribution and consumption is poor,which leads to the inaccuracy of the power anomaly analysis result,the anomaly analysis algorithm of electricity consumption data based on the least squares method and clustering is proposed.Firstly,the user electricity data is analyzed,and data features are normalized.Secondly,by analyzing the characteristics of the existing clustering algorithms,the K-means algorithm is used to analyze the data,and the data points are fitted by the least squares method,the threshold of the centrifugal point data is calculated,and the centrifugal points data is added to the noise set for isolation,which improves the efficiency of the K-means algorithm.Finally,the traditional K-means clustering algorithm is compared with the proposed algorithm,which proves that the proposed algorithm has achieved good results in both accuracy and false positive rate.
作者
张颖
王琳
王丽华
王飞
张苗
ZHANG Ying;WANG Lin;WANG Lihua;WANG Fei;ZHANG Miao(State Grid Hebei Electric Power Co.,Ltd.Shijiazhuang Power Supply Branch,Shijiazhuang 050000,China)
出处
《河北电力技术》
2019年第5期4-6,9,共4页
Hebei Electric Power
关键词
配用电数据
聚类算法
最小二乘法
K-MEANS算法
power distribution and consumption data
clustering algorithm
the least squares method
K-means algorithm
