Anomaly Detection Using Higher-Order Feature

Learning-based anomaly detection method is often subject to inaccuracies due to noise, small sample size, bad choice of parameter for the estimator, etc. We propose a novel method using higher-order feature, based on the sequence nonparametric test to assess the reliability of the estimation. The method allows an expert to discover informative features for separation of normal and attack instances. We performed experiments on the KDD Cup dataset. The results show that method reveals the nature of attacks. Application of the method yields a major improvement of detection accuracy.