Network Anomaly Detection using Data Mining Algorithms

Abstract

Anomaly detection is an area that is currently being explored. The purpose of this paper is to clarify a network anomaly detection framework that uses K-Means clustering and SVM classification to discover network characteristics in the NSLKDD dataset in order to reduce the false alarms rate, and further develop the positioning rate and identify the zero-day attacker. After preparing and testing the proposed data mining algorithm, the results show that the proposed method (K-Mean + SVM) has achieved a positive detection rate of (95.32%) and reduced the false alarm rate to (1, 1%), and reached (87.33%). ).