摘要
研究网络安全入侵准确检测问题。针对一些病毒和黑客攻击,传统入侵检测算法易出现检测正确率低和速度慢等问题,单独采用遗传算法不可以在较短的时间找到接近最优解,为了解决上述问题,提出了一种改进的遗传算法神经网络入侵检测系统。采用改进的遗传算法来优化BP神经网络权值,较好地与BP算法结合。采用已经训练好的BP神经网络对不匹配的可疑的入侵行文进行检测,并且能够识别检测出具体的网络入侵的类型。Matlab仿真结果表明,遗传算法与改进的BP神经网络有效结合在网络入侵检测中应用潜力很大,与较传统网络入侵检测系统模型相比,具有更好的入侵识别率和检测效果。
For some viruses and hacker attacks,the correct detecting rate of traditional intrusion detection algorithms is low and the speed is slow,therefore,using the genetic algorithm alone can not find the optimal solution in a short time.In order to solve these problems,an improved genetic algorithm of neural network intrusion detection system is proposed.The improved genetic algorithm is used to optimize the weights of BP neural network.The trained BP neural network is used to detect the suspected intrusion which does not match the language,and the specific network intrusion type can be identified.Matlab simulation results show that the effective combination of improved genetic algorithm with BP neural network has great potential in intrusion detection with a better recognition rate and detection rate.
出处
《计算机仿真》
CSCD
北大核心
2011年第6期173-176,共4页
Computer Simulation
关键词
遗传算法
算法优化
入侵检测
Genetic algorithm（GA）
Algorithm Optimization
Intrusion detection
