Detection of Outlier of Electric Power Data Based on Improved Fast Density Peak Clustering

Outlier detection is an important field in the application of electric power big data, and has an important application in depth analysis and mining of electric power data. The detection of outlier of electric power can avoid the fault, eliminate the source of interference of power quality and reduce the loss of power grid. This paper is set in the power big data, aiming at the characteristics of big data of electric power. In this paper, an improved fast density peak clustering is proposed, which introduces the idea of local outlier factor algorithm, and redefines the relative density and relative distance. The identification degree of cluster center with relatively small density in data is improved, and the accuracy of outlier detection is improved. The validity of the algorithm is proved by the simulation experiment on the outlier detection of the load data from an oil company.