A Novel Hybrid Consumer Behavior Classification Method for Demand Response

Ruihao Wang, Gengfeng Li, Yu Kou

Abstract


This paper presents a novel hybrid method of characterizing the residential consumer behaviors in the environment of demand response. Firstly, K-means algorithm is adopted to acquire a set of characteristic load curve. Distinguished from random initial point selecting in the process of traditional K-means algorithm, a hierarchical clustering method is used to construct a relatively better cluster center for reducing the computation time. Moreover, silhouette coefficient can be implemented to select the proper number of classification. Then, the improved support vector machines (SVM) is proposed to solve the classification problem of added residential consumer behaviors, making full use of previous cluster results. Finally, the numeric experiments demonstrate that the proposed is effective and efficient.

Keywords


Demand response, silhouette coefficient, hierarchical clustering, K-means algorithm, improved support vector machine


DOI
10.12783/dteees/appeec2018/23567

Full Text:

PDF

Refbacks

  • There are currently no refbacks.