MPCA: A modified principal component analysis method for anomaly detection

作者：Xu Guangxia^*; Liu Yu; Yao Zhangjun; Lin Mingwei

来源：International Journal of Advancements in Computing Technology, 2012, 4(16): 74-81.

DOI：10.4156/ijact.vol4.issue16.9

• Anomaly detection
• Anomaly detection systems
• Bayes decision rules
• Computer nodes
• Daily lives
• False negative rate
• False positive rates
• Network services
• Network systems
• Network technologies
• Principal component analysis method

With the rapid development of network technology and the ever-increasing of network services, people's production and daily life are increasingly dependent on network system. In this case, if some computer node in the network system emerges anomaly and does not support right services, we can not obtain right services from this computer node. In order to make the network system provide correct network services smoothly, we design an anomaly detection system to monitor the computer nodes in the network system. In this paper, we propose a modified principal component analysis method called MPCA, for anomaly detection. In order to reduce the false positive rate and the false negative rate, we introduce joint Gaussian distribution and Bayes Decision Rule to modify the traditional principal component analysis method. Experimental results show that our proposed MPCA significantly reduces the false positive rate and the false negative rate compared to traditional principal component analysis.

• 出版日期2012