Identification Method of Power Internet Attack Information Based on Machine Learning

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