Abnormal Traffic Detection Technology of Power IOT Terminal Based on PCA and OCSVM

The abnormal traffic detection technology based on the access boundary of the power Internet of Things (IOT) is a common method for detection and identification of terminal attack risk. Aiming at the problem that it is difficult to obtain attack samples and lack of effective prior knowledge in the abnormal traffic detection of power IOT terminals, the abnormal traffic detection technology of power IOT terminals based on PCA and OCSVM is proposed. Considering that the normal network traffic of each type of terminal has regular characteristics in the actual operation environment, model aggregation training can be conducted based on the normal operation traffic of the terminal. The technology extracts the key features of the traffic through PCA features, removes redundant information, and then constructs a detection model for the normal traffic behavior of the power IOT terminal based on OCSVM to detect and identify the abnormal traffic of the terminal. The simulation results show that the technology can effectively balance the classification accuracy of normal and abnormal samples of terminal traffic, and has better detection accuracy than some machine learning algorithms.