Research and application of One-class small hypersphere support vector machine for network anomaly detection

In recent years, machine learning technology often used as a recognition method of anomaly in anomaly detection. In this paper we have proposed a One-class small hypersphere support vector machine classifier (OCSHSVM) algorithm, which builds a learning classifier model via both normal and abnormal network traffic. This combination of normal and abnormal traffic for training model gives the better performance and generalization for proposed classifier Experimental results show that high detection rates and low false positive rates are achieves by our proposed approach. We have demonstrate proposed algorithm by using of KDD 1 and NSL-KDD 2 dataset.