Research on Intrusion Detection Based on KPCA-BP Neural Network
Ding Hongwei, Liang Wan · 2018
In view of the traditional BP neural network, high-dimensional complex data is prone to slow detection rate and low accuracy in network intrusion detection. To reduce data dimension and improve BP neural network performance, an intrusion detection method of KPCA-BP neural network is proposed. Firstly, the KPCA's good dimensionality reduction capability is used to reduce the dimension of network data. Then, by changing the initialization initial value method and loss function of traditional BP neural network, the learning performance of BP neural network is improved, and the learning effect of improved BP neural network is better. Experiments show that the KPCA-BP based intrusion detection method proposed in this paper has a better improvement effect on detection rate and accuracy.