Research on Network Intrusion Detection Based on GRU_SVM

Abstract:
With the increasingly in-depth integration of network and society, network security has become a hot issue of concern. Intrusion detection technology is one of the important guarantees of network security operation. In order to solve the problem of low detection rate in intrusion detection, especially for small sample attack types, a method based on GRU is proposed - Intrusion detection method based on SVM model. This method constructs the training model based on tensorflow learning framework, and preprocesses the intrusion detection dataset through pandas, numpy and other function libraries. Using the excellent feature extraction and learning ability based on gated cyclic unit network (GRU), combined with the easy fit of multilayer perceptron (MLP), the preprocessed data are feature learned and its vector is reduced to low dimension. Using the good classification ability of support vector machine (SVM) in low dimensional features, the dimensionality reduced attack behavior is detected and classified, so as to improve the detection rate of the detection system.