Аннотации:
To solve the problem of large recognition errors in traditional attack information identification methods, we propose a machine learning (ML)-based identification method for electric power Internet attack information. Based on the Internet attack information, an Internet attack information model is constructed, the identification principle of the power Internet attack information is analysed based on ML, hash fixing is conducted to ensure that the same attack information will be assigned to the same thread and that the deviation generated by noise can be avoided so that the real-time lossless processing of the power Internet attack information can be ensured. The vulnerability adjacency matrix is constructed, and the vulnerability is quantitatively evaluated to complete the design of the optimal identification scheme for power Internet attack information. The experimental results show that the identification accuracy of the method сan reach 98%, which can effectively reduce the risk of power Internet network attacks and ensure the safe and stable operation of the network.
