Network invasion abnormity detection method

Abstract

For an invasion detection model construction problem, the present invention provides a new network invasion abnormity detection method of supervising the non-linear feature extraction and a regularization random weight neural network (RRWNN). A kernel partial least square (KPLS) algorithm is used to process the colinearity of the input features and the complicated nonlinear mapping. The extracted potential features are inputted to an RRWNN algorithm to construct an invasion detection model possessing a higher learning speed and a better generalization performance. A global optimization strategy is adopted to select the modeling parameters of a KPLS-RRWNN-based invasion detection model, and the simulation based on the KDD99 data indicates the validity of the method.

Description

technical field [0001] The invention relates to network security technology, in particular to a network intrusion abnormality detection method. Background technique [0002] With the wireless and mobile network, and the gradual deepening and promotion of Industry 4.0, the information security situation is becoming more and more serious, and the security protection of key infrastructure related to national security needs to be paid more attention. Network intrusion detection refers to discovering behaviors that violate security policies or endanger system security in the system by collecting information such as operating systems, system programs, applications, and network packets. Existing network intrusion anomaly detection techniques include statistical analysis, pattern prediction, neural networks, genetic algorithms, sequence matching and learning, immune systems, norm-based, data mining, integrity checking, and Bayesian techniques. [0003] In this context, it is very i...

AI Technical Summary

Problems solved by technology

However, it cannot guarantee that the constructed random weight neural network (RWNN) model has universal approximation ability (in the sense of probability 1), which is mainly due to the way in which the input weights and biases are randomly given (see "Cao, F.L, Wang, D.H, Zhu, H.: An iterative learning algorithm forfeedforward neural networks with random weights. Information Sciences1(9), 546-557(2016).”)

Images