Principal component analysis and support vector machine based anomaly detection
Kong Qiang · Jisuanji yingyong yanjiu · 2009
To improve the efficiency of anomaly detection,this paper proposed one principal component analysis (PCA) and support vector machine (SVM) based intrusion detection method. Employed PCA to reduce intrusion data,SVM train the reduced data,and constructed one anomaly detection model to detect the test data. Kddcup'99 data based experiments,40 features is reduced to 15,and 22 features is reduced to 5 features. Experimental results on these data show presented method have strong general ability and spend less training time and test time compared with using all features.