摘要： 为解决传统入侵检测算法存在的高漏报率及高误报率问题，结合BP神经网络算法的优点，提出一种采用遗传算法来优化BP神经网络算法的入侵检测算法。该算法通过遗传算法找到BP神经网络的最适合权值，采用优化的BP神经网络对网络入侵数据进行学习和检测，解决直接使用BP学习造成的训练样本数量过大而难以收敛的问题，同时缩短样本训练时间，提高BP神经网络分类正确率。仿真实验结果表明，与传统网络入侵检测算法相比，该算法的训练样本时间更短，具有较好的识别率和检测率。

关键词: BP神经网络, 算法优化, 入侵检测, 漏报率, 误报率

Abstract: Traditional intrusion detection algorithm for the existence of a relatively high false negative rate and high false alarm rate defects, based on the advantages of BP neural network algorithm, this paper proposes a new algorithm that uses genetic algorithm to optimize BP neural network algorithm for the intrusion detection. The genetic algorithm finds the most appropriate BP neural network weights, and uses the optimized BP neural network learning and network intrusion detection data. Algorithm effectively improves the classification accuracy of BP neural network. Matlab simulation experimental results show that the training samples of the proposed detection algorithm have less time, with better recognition rate and detection rate, compared with traditional intrusion detection algorithm.

Key words: BP neural network, algorithm optimization, intrusion detection, false negative rate, false alarm rate

中图分类号: 

• TP319