Electricity User Consumption Feature Selection and Behavior Portraying

The stable operation of China's AC/DC bulk power grid and the accommodation of renewable energy put forward high requirements on the demand response. Electricity user portraying is of great significance to the implementation of the demand response. A method for electricity user consumption feature selection and behavior portraying is proposed in this paper. Firstly, the optimal cluster number is obtained automatically by constructing a novel clustering return index taking into account the degree of aggregation and separation. The K-means clustering algorithm is used for user consumption samples clustering. Then, the maximal relevance and minimal redundancy (mRMR) criterion is applied to the selection of electricity user consumption features for balancing effectiveness and simplicity. A high-quality feature set is obtained by using the enumeration method. Finally, the high-quality features are quantified by using the scoring method, then the radar map and the histogram are used to show the electricity user consumption behavior portraying. The results of an example show that the proposed method is effective.