Application of an Improved multi-layer BP Neural Network Algorithm in Intrusion Detection

this paper analyzes the shortcomings of traditional multilayer BP neural network in intrusion detection, and gives an optimized multilayer BP neural network algorithm based on matrix, which has good readability, higher recognition rates and the faster program execution speed, etc. On the basis of this, the improved algorithm, combined with principal component analysis and other methods, is applied to intrusion detection. Theoretical analysis and experimental data show that the proposed algorithm is more concise and it can detect intrusions more quickly and efficiently, compared with the traditional multi-layer neural networks.