一种基于融合孤立点挖掘的入侵检测技术

An Intrusion Detection Method Based on Outlier Detection

将融合概念引入孤立点挖掘中,对基于相似度和的孤立点挖掘算法进行融合,提出一种基于投票机制的融合孤立点挖掘算法VoteSimiOut,并将其应用于入侵检测中。采用编码映射方法对符号型数据进行处理,并利用主成分分析来实现对编码映射后扩展的属性进行降维。详细阐述了具体实现方案,并通过仿真实验验证了该方法的可行性。

Brings the concept of ensemble into outlier detection, outlier mining algorithm of similar coefficient sum be ensembled. Proposes an anomaly detection method based on voting mechanism （VoteSimiOut） and applies it into intrusion detection. Transforms the character feature into numerical value by code mapping and uses Principal Components Analysis （PCA） to reduce dimeusion.