入侵检测系统在训练过程中需要大量有标识的监督数据进行学习,不利于其应用和推广.为了解决该问题,提出了一种基于主成分分析的无监督异常检测方法,在最小均方误差原则下学习样本的主要特征,经过压缩和还原的互逆过程后能最大限度地复制样本信息,从而根据均方误差的差异检测出异常信息.构建的仿真系统经过实验证明,基于主成分分析的无监督异常检测方法能够在无需专家前期参与的情况下检测出入侵,实验结果验证了其有效性.-Intrusion Detection System in the training process

K-means Algorithm Based on Particle Swarm Optimization Algorithm for Anomaly Intrusion Detection

The title of my paper is An overview of anomaly based intrusion detection systems .-The title of my paper is An overview of anomaly based intrusion detection systems .