论文部分内容阅读
针对大规模网络环境下的入侵检测系统需要处理的网络数据含有大量的冗余与噪音的问题,设计了一种基于轻量级人工免疫计算的混合入侵检测方法.利用最小信息熵离散化算法预处理检测数据,根据主元分析算法(PCA)进行特征提取,通过提取特征矩阵降低数据维度;设计了基于否定选择算法的在线检测,对于未知的或者大规模的连接则提取其特征并实现基于人工免疫计算的入侵检测.最后利用进化能力的异常检测器进行训练和检测,并将提取的异常特征模式加入到快速匹配的数据库来及时地更新数据库.仿真实验表明算法能够提高混合检测器系统的检测效率,同时检测速度能够满足实时性的要求.
In order to solve the problem of large amount of redundancy and noise in network data to be dealt with in intrusion detection system under large-scale network environment, a hybrid intrusion detection method based on lightweight artificial immune algorithm is designed. By using the minimum information entropy discretization algorithm Processing the detection data, extracting features by principal component analysis (PCA) algorithm and reducing the data dimension by extracting feature matrix; designing on-line detection based on negative selection algorithm, extracting features for unknown or large-scale connections and implementing artificial detection based on artificial Immune computing intrusion detection.At the end of the evolutionary ability of the anomaly detector for training and testing, and extract the abnormal feature patterns to quickly match the database to update the database in time.Experimental results show that the algorithm can improve the detection of hybrid detector system Efficiency, while testing speed to meet the real-time requirements.