Application of Parallel Clustering Algorithm Based on R in Power Customer Classification

The power load curve is describing the behavior of electricity consumption by customers. The curve characteristic information is obtained by mining and analyzing the power load curve, which can be used as the classification basis for power customers. Based on the massive characteristics of power load curve, the K-means clustering algorithm and Spark-based parallelization method are studied firstly, and on the basis of this, the parallel transformation of k-means algorithm is carried out. Then the preprocessing steps of the power load curve data and the detailed flow of the mining analysis are proposed, and the R parallelization algorithm is applied to the mining analysis of the actual power load curve data, through which the classification of electricity customers are achieved and verified. Finally, the R-based parallel clustering algorithm is briefly summarized.