The construction of electric power user profiles is often constrained by limited data dimensions, superficial feature extraction, and static tag libraries. To address these challenges, this study proposes a novel integrated cluster‐driven framework. This framework commences with an integrated data processing operation that ensures the quality and fusion of multi‐source data. On this basis, develop a dynamic user behavior tagging system that incorporates multi‐dimensional feature extraction and an adaptive update mechanism. An enhanced K‐means clustering algorithm, adopting the Mahalanobis distance to capture feature correlations, serves as the core of our methodology, enabling more accurate aggregation of user behaviors. The resultant feature sets are visualized through radar charts, providing an intuitive representation of distinct consumption characteristics. Experimental results demonstrate that the proposed framework effectively discriminates between diverse user consumption patterns, offering a valuable theoretical basis for refined management and optimal resource allocation in power enterprises. © 2026 Institute of Electrical Engineers of Japan and Wiley Periodicals LLC.
Xuepeng et al. (Wed,) studied this question.