作者: Michał Choraś , Łukasz Saganowski , Rafał Renk , Witold Hołubowicz
DOI: 10.1111/J.1468-0394.2010.00576.X
关键词: Discrete wavelet transform 、 Anomaly detection 、 Data mining 、 Network security 、 Anomaly-based intrusion detection system 、 Intrusion detection system 、 Signal processing 、 Computer science 、 False positive paradox 、 Matching pursuit
摘要: In this paper, a framework for recognizing network traffic in order to detect anomalies is proposed. We propose combine and correlate parameters from different layers 0-day attacks reduce false positives. Moreover, we statistical signal-based features. The major contribution of paper novel security based on the correlation approach as well new algorithm intrusion detection basis Matching Pursuit (MP) algorithm. As our best knowledge, are first use MP anomaly computer networks. presented experiments, proved that solution gives better results than discrete wavelet transform. © 2012 Wiley Periodicals, Inc.