Insertion Detection System Employing Neural Network MLP and Detection Trees Using Different Techniques

Authors

M. H. Al-Mashagbeh, Cyber Security Department, Faculty of Information Technology, Zarqa University, Zarqa, JordanFollow
W. Salameh, Computer Science Department, Faculty of Computing Sciences, Princess Sumaya University for Technology, Amman, JordanFollow
A. B. Alamareen, Computer Science Department, Faculty of Information Technology, Zarqa University, Zarqa, JordanFollow
S. Abu asal, Software Engineering Department, Faculty of Information Technology, Zarqa University, Zarqa, JordanFollow

Author Country (or Countries)

Jordan

Abstract

by addressing intruder attacks, network security experts work to maintain services available at all times. The Intrusion Detection System (IDS) is one of the available mechanisms for detecting and classifying any abnormal behavior. As a result, the IDS must always be up to date with the most recent intruder attack signatures to maintain the confidentiality, integrity, and availability of the services. This paper shows how the NSL-KDD dataset may be used to test and evaluate various Machine Learning techniques. It focuses mostly on the NLS-KDD pre-processing step to create an acceptable and balanced experimental data set to improve accuracy and minimize false positives. For this study, the approaches J48 and MLP were employed. The Decision Trees classifier has been demonstrated to have the highest accuracy rate for detecting and categorizing all NSL-KDD dataset attacks.

Digital Object Identifier (DOI)

https://dx.doi.org/10.18576/jsap/130112

Recommended Citation

H. Al-Mashagbeh, M.; Salameh, W.; B. Alamareen, A.; and Abu asal, S. (2024) "Insertion Detection System Employing Neural Network MLP and Detection Trees Using Different Techniques," Journal of Statistics Applications & Probability: Vol. 13: Iss. 1, Article 12.
DOI: https://dx.doi.org/10.18576/jsap/130112
Available at: https://digitalcommons.aaru.edu.jo/jsap/vol13/iss1/12