Power distribution communication network security situation awareness and abnormal intrusion detection method

Abstract

The invention discloses a power distribution communication network security situation awareness and abnormal intrusion detection method, which comprises the steps of performing data preprocessing on original network behavior data in network key nodes, and integrating the data into a standardized training data set; on the basis of a random forest algorithm, key feature indexes influencing the abnormal state of the network are extracted from the standardized training data set; constructing a feature layer forest according to the standardized training data set, training the feature layer forest in combination with the key feature indexes, and calculating a connection weight; establishing a network anomaly detection model according to the connection weight, and identifying a network attack type; according to the method, the training speed of wide forest learning algorithm modeling is increased, and the complexity of learning tasks is reduced; meanwhile, the model complexity is reduced, the learning convergence speed is increased, and the detection accuracy is improved.

Description

technical field [0001] The invention relates to the technical field of intrusion detection of communication networks, in particular to a security situation awareness and abnormal intrusion detection method of a power distribution communication network. Background technique [0002] With the comprehensive and rapid development of information technology and network technology, the operational reliability and stability of power distribution communication network have been further improved. The application of information technology not only improves the working efficiency of the power system, but also simplifies the work process, but it also brings corresponding network security protection problems. Once the power distribution communication network is affected by the outside world, destroyed or interfered, it cannot work normally. It will seriously affect the production and life of the people and cause significant economic losses. At present, there are still many problems in th...

AI Technical Summary

Problems solved by technology

Traditional network intrusion behavior detection methods take a long time to detect and have low detection efficiency; network intrusion behavior detection methods based on machine learning, such as methods based on cluster analysis and support vector machines, have high intrusion feature data dimensions and data between different intrusion categories. When the difference is small, many intrusions cannot be accurately detected, and there is also a high false alarm rate; although the accuracy of network intrusion behavior detection methods based on neural networks has been significantly improved, the training speed is slow, the classification accuracy is low, and the implementation is complicated. high degree

Images