Cluster Analysis of Power User Loads Based on KPCA and K-means++

In view of China's fiercely competitive power market, more power companies want to study the power consumption habits of power users to provide more personalized services. This paper proposes a power user load clustering method based on KPCA and K-means++. After the user load data is dimensionally reduced by KPCA, K-means++ is used for clustering processing. The experiment uses the 2018 power user load data in a region in northeastern China to simulate. The results show that the proposed method has improved effectiveness and stability compared with other traditional clustering methods.