Anti-electricity-stealing identification method based on bidirectional long-short-term memory network and sliding window input

The invention discloses an anti-electricity-stealing identification method based on a bidirectional long-short-term memory network and sliding window input, and the method comprises the steps: collecting the electricity consumption information of an electricity-stealing user and a normal user from an electricity consumption information collection system, and manually adding a label for judging whether there is an electricity-stealing behavior or not to the electricity consumption information; adopting a sliding window input mode to input electric quantity data in a continuous period of timeincluding a target day and a label of the target day, and learning input electric quantity sequence features through a bidirectional long-short-term memory network; and forming an electricity-stealing intelligent identification model, analyzing the electricity consumption data of the user, and judging whether the user has electricity-stealing suspicion or not. According to the invention, label processing is carried out on t