In order to achieve electricity theft detection, this paper proposes a methd through the analysis of electricity consumption behavior of users, which is based on non-instrusive load monitring. In this method, the commonly-used process is carried out, including the load detection, feature extraction and meanshift clustering, in order to obtain the features of each type load inner the home. Thus, the dataset of load is built which consists of the non-electrical features, such as the time point of load turing on and off, the length of load works, and so on. Thereby, the electricity theft detection model can be built for load cluster and probability prediction. Meanwhile, according to the electrical electricity theft detection model, the predition is performed by the information, including the load event, the length of load work and the energy consumption. The bayes theory is then introduced to infer whether its electrical consumption behavior is normal or not. Finally, the experiments are carried out by using the real information from the smart meter. The results show that the proposed method can provide the basis and support for the electriity theft detection.
