AN EFFIECIENT CYBER ATTACK DETECTION USING MACHINE LEARNING TECHNIQUES

COMPUTER ENGINEERING

A new generation of technology has evolved which allows for the transmission of data between a utility company and its customers in real-time through new technologies such as Advanced Metering Infrastructure. The security and privacy of smart grid systems, which combine smart and legacy information and operational technologies, have grown in concern. We propose an information attack detection model for the smart grid based on XGBoost. It uses a modified k-means-smote oversampling method to obtain a balanced power data set, which solves the problem of data imbalance causing high false-positive rates in network attack detection. On the basis of oversampling data, feature selection is performed to reduce the dimension of the data. This will shorten the model training time, and accelerate the response speed of the network attack detection model. Finally, construct an XGBoost classifier model to identify several network attack modes in the data set. The paper studies machine learning models and proves that the network attack detection model improves the detection accuracy of smart grid information attacks significantly.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.