An Approach to Identification Model for Electric Event Based on Clustering Analysis

Identifying user electric events is of significant importance for uncovering patterns in user electric consumption behavior and enhancing the level of energy efficiency management on the user side. To promptly and effectively detect electric events embedded in the electric data of the user, this paper introduces a cluster-based electric event identification model designed based on the electric current state sequence dataset. Based on extracting features from the electric current state sequence, this model treats the identification of electric events in the current sequence as a clustering partition problem utilizing the feature set derived from the electric current state sequence. To assess the model’s efficacy, two metrics were utilized: the silhouette coefficient and precision, to evaluate its performance. The experiments demonstrate that compared to the identification model for the electric event of the user based on k-means clustering, SOM clustering, and FCM clustering, the identification model for the electric event of the user based on hierarchical clustering is more effective in identifying electric events.