针对互联网规模的不断扩大,网络服务更容易向入侵者和攻击者暴露信息,且攻击手段日趋复杂,提出了简化粒子群优化(SPSO)结合自组织映射(SOM)的网络入侵检测方法.根据特征判别力,使用PCA方法进行选择特征,生成非相关性特征过滤数据集噪声和低方差值特征.通过SOM与高斯混合模型(GAMM)混合方法来模拟正常模式与异常模式,测量每个网络单元的激活概率以检测所有高频率攻击的精确值,并运用概率SOM均值对特征空间进行分类,在此过程中,运用简化粒子群优化(SPSO)算法从分类搜索当前解的邻域内找到更优的解.基于KDDCUP99数据集搭建仿真测试平台,实验结果表明,提出的方法对常见的网络攻击表现出了良好的性能,具有更高的入侵检测准确率(ACC).%For the issue of growing size of the Internet,information of network services is more likely to be exposed to intruders and attackers,and network attacks are becoming more and more complex.A method of network intrusion detection using simplified particle swarm optimization (SPSO) and self organizing mapping (SOM) is proposed.According to the feature discrimination force,the PCA method is used to select the feature,and the non-correlation feature filtering data set noise and low variance feature.The normal mode and anomaly modes are simulated by SOM and Gaussian mixture model (GAMM).The scheme allows the measurement of the activation probability of each network element to detect the exact value of all high frequency attacks and classifies the feature space using the probability SOM mean.In this process,a simplified particle swarm optimization (SPSO) algorithm is used to find a better solution in the neighborhood of the current solution.Based on the KDDCUP99 data set,the experimental results show that the proposed method has a good performance and a higher intrusion detection accuracy (ACC) for common network attacks.
展开▼
