Clustering Analysis of Power Grid Data Based on Dynamic Hierarchical K-Modes

With the development of power industry and the continuous improvement of power metering system, the demand side power consumption characteristics of power grid are diversified. How to tap the characteristics of users’ electricity consumption behavior, so as to promote the marketization of electricity price, has become a concern. Firstly, the K-Modes algorithm in clustering analysis is introduced, and a dynamic hierarchical K-Modes algorithm is proposed to deal with generic data and give a reasonable H value; Secondly, the data processing method of curve data difference and generic transformation is proposed to better reflect the user curve shape; Finally, the dynamic hierarchical K - modes algorithm is used to get good classification results on simulated data.