A Security Posture Prediction Method for Power Networks Based on Extreme Learning Machine
Jun Ma, Qian Xu, Miaojie Zou, Jiong You · 2023
To address the problems of single information source and lack of real-time of the existing network security posture prediction, an extreme learning machine based power network security posture prediction model is proposed by examining the characteristics of network security posture change in order to accurately grasp the network security development posture. First, the security posture value data set is constructed using the sliding window method, and the training sample set is used to train the limit learning machine neural network to portray the backward and forward dependencies of the security posture at different moments and predict the network security posture at the next moment. The effectiveness of the prediction model is verified by the dataset based on the on-site security monitoring data of power enterprises. The simulation experiments show that the method can predict the network security posture more accurately, which verifies the feasibility and effectiveness of the method proposed in the paper in network security posture prediction.