Load curve clustering has important applications in load forecasting and demand-side response. At present, the load data is becoming more and more quantized and multi-dimensional, and it needs to be reduced in dimensionality. However, the existing dimensionality reduction will cause a certain degree of loss in the curve information. Therefore, a multi-dimensional scaling (MDS) Daily load curve clustering method for dimension reduction. First, the MDS algorithm is used to perform dimensionality reduction on the collected load data, then the CRITIC-entropy weight method is used as the weight configuration method of the dimensionality reduction index, and finally the optimal cluster number is selected by the weighted modified contour coefficient to weight the European Distance K-means algorithm for clustering. A numerical example shows that this method can retain the original curve information to the greatest extent, and has advantages in both clustering accuracy and running time.
