Research on User Classification Method of Electricity Market Based on K-means

Abstract

In any industry, categorizing users and managing them in a targeted manner has always been an effective method. However, with the development of the times, there are more and more users, and more and more data can be collected, which has caused a certain degree of difficulty in user classification. User classification has always been one of the key researches in all walks of life. As one of the classic algorithms in unsupervised learning, the K-means clustering method has a good clustering effect on unlabeled data. According to the characteristics of the electricity consumption data of electricity market users, this paper proposes a user classification model based on K-means clustering algorithms. Firstly, it selects the median and $3\sigma$ criterion to complete the missing values and clean up the abnormal values of the original electricity consumption data. Then use principal component analysis to reduce the dimensionality of the data, and finally use the K-means clustering algorithm to cluster the processed data. Through clustering, the users are divided into different clusters, each cluster is equivalent to a user group, and at the same time extract, the typical users of this user group are the representatives of this user category. The model is verified through the analysis of calculation examples, and the experimental results show that the classification of power users based on the K-means clustering algorithm is very effective.