as an new kind of signal process method, independent component analysis ( ica ) has been applied on many domains such as finance, biomedicine, image manipulation and so on . we apply the model of ica on intrusion detection system and propose three new ids algorithmes which combine the idea of k _ mean algorithm, peseduo bayesian estimator and support vector machine with ica in this paper . in the space assured by ica, the three algorithms can prove the performance of classfier 本文将独立分量分析模型引入到入侵检测系统中,提出了将独立分量分析模型同k-均值算法、伪贝叶斯估计、支撑矢量机相结合的三种入侵检测算法。通过将样本投影到由独立分量分析所确定的子空间,这三种算法弥补了原有算法的弱点,大大提高了入侵检测系统的性能,并对各算法的进一步研究和发展提出了自己的看法。