Household User Behavior Analysis based on Power Data

With the popularization of smart grid application and the rapid development of big data technology, a large amount of user power consumption data is recorded and stored. making user power consumption behavior analysis the key content of demand side management. The user behavior state is identified by the variation of power amplitude of power load based on hidden Markov model. Users are mapped to potential feature space by means of price elasticity information regularization matrix decomposition. Further, K-means clustering algorithm is adopted to segment users. The strategy can improve the rationality of user power consumption behavior clustering, more effectively grasp the form of user power consumption load, reduce the peak load of power and balance power supply and demand, which is of great significance to the demand side management of smart grid.