入侵数据特征约简方法综述

Overview on Feature Reduction Approach of Intrusion Data

文中主要研究了目前比较流行的入侵数据特征约简方法,包括主成分分析法(PCA)、独立成分分析法(ICA)、基于粗糙集的属性约简方法以及基于SVM和神经网络实验的特征选择方法。分析比较了各种方法的优缺点、适用场合及其各自在入侵数据特征约简中发挥的作用。

This paper studies the popular methods for feature reduction of the intrusion data, including the principal component analysis （PCA）,the independent component analysis （ICA）, attribute reduction based on rough set, and the feature selection based on the experiments of SVM andneural network. It also analyses the advantages and disadvantages of various methods, the applications and the roles they play respectively in feature reduction of the intrusion data.