Multivariate correlation analysis technique based on Euclidean distance map for network traffic characterization

摘要: The quality of feature has significant impact on the performance detection techniques used for Denial-of-Service (DoS) attack. features that fail to provide accurate characterization network traffic records make suffer from low accuracy in detection. Although researches have been conducted and attempted overcome this problem, there are some constraints these works. In paper, we propose a technique based Euclidean Distance Map (EDM) optimal extraction. proposed runs analysis original space (first-order statistics) extracts multivariate correlations between first-order statistics. extracted correlations, namely second-order statistics, preserve discriminative information characterizations records, can be high-quality potential DoS attack effectiveness is evaluated using KDD CUP 99 dataset experimental shows encouraging results.