HIDS:DC-ADT : An Effective Hybrid Intrusion Detection System based on Data Correlation and Adaboost basedDecision Tree classifier

Main Article Content

Ali Raeeyat
Hedieh Sajedi


Due to the rapid development of computer networks, intrusions and attacks into these networks have grown, and occur in various ways. Thus, usually an intrusion detection system can play an important role in security protection and intruders’ accessibility to network prevention. In this paper, a new hybrid approach, which is called HIDS:DC-ADT, is used to design proposed detection engine. In the proposed intrusion detection system, the anomaly detection engine is responsible to detect new and unknown attacks and the misuse detection engine is responsible to protect anomaly detection system.Through this, it is assured that collected data and patterns be safe for anomaly detection system.In the intrusion anomaly detection using statistical correlation method that is of data correlation methods, normal behavior of network is analyzed statistically by KDD-Cup99 data-set. Further, the Data Correlation Graph (DCG) has been proposed to show behaviour’sdeviation of normal behavior. In misuse detection, Principal Components Analysis (PCA) is used to dimensionality reduction. More, a new classification method by Adaboost algorithm using base classifier of decision tree C4.5 has been introduced for classification. Simulation results show that this hybridsystem can reach a competitive accuracy and efficiency.

Keywords:HybridIntrusion Detection System; Data Correlation; Data Correlation Graph; Adaboost Algorithm, Decision Tree; Principle Component Analysis.


Download data is not yet available.

Article Details