Research on Load Clustering Based on Singular Value Decomposition and K-means Clustering Algorithm
Jiao Wang, Kaiyan Wang, Rong Jia, Xiaofei Chen · 2020 Asia Energy and Electrical Engineering Symposium (AEEES) · 2020
Considering the shortcomings of existing clustering algorithms in clustering quality, this paper proposes a load clustering research based on singular value decomposition and K-means clustering algorithm. First, eight characteristic indexes of load are extracted, and the singular value is used to decompose the load characteristics of the user side. The solved singular value reflects the importance of this type of load characteristic. The load characteristics corresponding to the data with large singular value are taken as the main load characteristics to complete the dimensionality reduction of the data. Then, the clustering evaluation index SSE is used to compare the effects of direct clustering of load characteristics and clustering after singular value decomposition. The results show that the proposed method has better clustering effect on load characteristics. Finally, K-means clustering algorithm is used to cluster the load characteristics.