Industrial control Internet of Things abnormal behavior detection method and system based on comparative learning

The invention relates to an industrial control Internet of Things abnormal behavior detection method based on comparative learning, which is characterized in that a comparative learning algorithm is used for improvement, on the basis of the design of adding an auxiliary label to a sample, a hierarchical comparative loss function is introduced to improve learning of a model on feature information, and network training under multi-dimensional loss analysis is realized in combination with cross entropy loss. A network abnormal behavior detection model corresponding to the target industrial control Internet of Things is obtained, and the network abnormal behavior of the industrial control Internet of Things is accurately detected; a corresponding system is further designed, application throughout the whole scheme design is achieved through modular design, efficient execution of the detection method in practical application is achieved, anomaly detection of the industrial control Internet of Things is accurately a