Classification Method For User Electricity Consumption Behavior Based on Power Big Data

School of Information, Shenyang Institute of Engineering, Shenyang, Liaoning, China

^* Corresponding author: yushun@sie.edu.cn

Abstract

According to the insufficiency of existing methods of classification on the behavior of using electricity, this paper proposes a classification method for behavior of user using electricity based on deep convolutional neural networks(DCNN). The method which we proposed can classify the behavior of using electricity, then can support efficiency data for management and optimization of power system. This paper implemented classification mainly on user portrait and the big data of power, based on modeling the electricity using of users’, we used deep convolutional neural networks as our fundamental classification method, and made several improvements in connection with the model of using electricity we had built. The experiments shows that the accuracy of classification method which we proposed could reach 90% and more as a whole, which verified the effectiveness and feasibility of the model and method , and this result can also provide references for the distribution adjustment of our existing power system.