| 王博,罗超.基于改进的BP神经网络的入侵检测研究[J].井冈山大学自然版,2011,(4):66-70 |
| 基于改进的BP神经网络的入侵检测研究 |
| RESEARCH ON INTRUSION DETECTION BASED ON IMPROVED BP NEURAL NETWORK |
| |
| DOI： |
| 中文关键词: BP 神经网络 入侵检测 局部最小 梯度下降 |
| 英文关键词: BP neural network intrusion detect local minimum gradient descent |
| 基金项目:安徽省高校自然科学研究项目(KJ2011B186) |
|
| 摘要点击次数: 1799 |
| 全文下载次数: 0 |
| 中文摘要: |
| 本文提出了引入可调因子和遍历局部最小并逃逸的方法,以解决标准BP算法中误差曲面过于平坦导致迭代次数增加、易陷入局部最小的缺点,并将此算法应用到网络入侵检测系统,对五类入侵行为进行检测。实验结果表明,改进后的BP算法大大缩短了系统响应时间并降低了检测系统的漏检率和误检率,极大改善了入侵检测系统的性能。 |
| 英文摘要: |
| In order to resolve two shortcomings in standard BP algorithm,an improved BP algorithm was proposed.In the improved algorithm,an adjustable factor and the methods of escaping from the local optimization were introduced.The algorithm was applied to network intrusion detection system.The experiment and analysis were shown that the system response time was shortened,and the missing rate and false detection rate of the intrusion system were reduced,the performance of the intrusion detection system was greatly i... |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
