Analysis of Power Utilization Characteristics Based on Clustering and Classification
The analysis of power utilization characteristics has close relations to some important decisions such as electricity pricing and distribution network planning. The load curves can reflect the actual characteristics of power utilization. The analysis of power utilization characteristics based on load curves can make related work more reasonable. This paper presents a method based on a combination of clustering and classification to analyze the power utilization characteristics of distribution network customers. Two steps are involved in the proposed method, they are k-means clustering to form typical load curves and a decision tree model based on the clustering result for customer classification. C4. 5 classification algorithm is used to classify customers of distribution network. Load classification accuracy is increased by selecting the optimal decision tree attributes. This method is applied to analyzing the load curves of distribution network customers. The result can reflect the actual power utilization characteristics and provides an effective pricing basis for managers.
Power Utilization Characteristics Load Curves K-Means Clustering Decision Tree C4. 5
Kong Yinghui An Jing Liu Yunfeng CHE Linlin
School of Electrical and Electronic Engineering, North China Electric Power University,Shenyang P.R.China, 071003
国际会议
天津
英文
2007-10-20（万方平台首次上网日期，不代表论文的发表时间）