Anomaly detection based on contiguous expert voting algorithm

Abstract

Malicious intrusion is the behavior that threats a large number of computers; therefore, recent research has focused on devising new techniques to detect and control internet intrusion with high efficiency and low cost. Unfortunately some anomaly detection system (ADS) over machine learning may get some false alarms if the results of machine learning cannot cover all the normal or abnormal data. In this paper, to solve this problem, we introduce a new approach for anomaly detection using contiguous expert voting algorithm (CEVS). At first, we present our framework of the anomaly detection system, and then we define a new algorithm based on data mining, at last we will use this algorithm to detect the internet anomaly and report our experimental result. The results show that the proposed approach can improve the detection performance of the ADS, where traditional anomaly detection system is used.