Network Security Analysis Method for Power System

Network security have been widely concerned by the academic community on the power system. Aiming at the problem that edge servers in power network environment are vulnerable to attacks, we presented a new machine learning network security analysis method. Based on intrusion detection network attack benchmark dataset, a proposed hybrid feature selection method is proposed by using similarity measurement method and random forest. A security analysis model is built by using ensemble algorithm. The main hyper parameters are optimized for the security analysis model. The comparison results show that the new method can effectively and automatically classify network attacks for network traffic.