Research on Data Mining-based Intrusion Feature Selection and Construction
基于数据挖掘的入侵特征选择与构造的新方法*

Intrusion detection is a classification issue, performance of classification model is determined by feature selection. Computer security dont be considered when designing computer system, there is no standard auditing mechanism and specific data format for intrusion detection analyst. This paper discusses feature selection method through extending basic data mining algorithm analyzing data source, constructing new features to increase accurate intrusion detection rate through comparing mined normal patterns and anomaly patterns.