Statistical, Forecasting and Metaheuristic Techniques For Network Anomaly Detection
; Pena, E. P.
; Carvalho, L. C.
Rodrigues, J. R.
; Proença, M. P.
Statistical, Forecasting and Metaheuristic Techniques For Network Anomaly Detection, Proc ACM Symp. on Appl. Computing, Salamanca, Spain, Vol. USB, pp. 1 - 7, April, 2015.
Digital Object Identifier: 0
Traffic monitoring is an arduous task and requires mech- anisms to proactively detect anomalous events that may harm the proper functioning of computer networks. Since the emergence of network management, several approaches have been developed to address this issue. In this paper, we examine and compare three methods used for anomaly detection: the statistical procedure Principal Component Analysis, the Ant Colony Optimization metaheuristic and the AutoRegressive Integrated Moving Average forecasting method. Experimental results on traffic collected at the backbone of a University network demonstrate high confi- dence in detection accuracy.