Enhance Network Intrusion Detection System Using Bee Algorithm

  • Soukaena H. Hashem University of Technology, Computer Sciences, Baghdad, Iraq.
  • Shatha Habeeb University of Technology, Computer Sciences, Baghdad, Iraq.
  • Besmah M. Khalil Al-Rafidain University College, Baghdad, Iraq.
Keywords: IDS, feature reduction, KDD training data, bee’s algorithm.

Abstract

Intrusion detection systems have sequential steps begin with selecting training and testing dataset, the preprocessing dataset, selecting most important features, and finally constructing the most reliable classifier. This research focuses on building a reliable Network Intrusion Detection System (NIDS) to detect traditional and modern attacks with minimum number of features. The proposal creates dataset depending on KDD. The proposal will inject KDD with new sessions to represent most modern attacks. This update requires adding new features for the dataset, since these features are critical to detect these modern attacks. The proposal considers updated dataset without any assumptions says that the dataset is idealism, so there are preprocessing steps to be done to make it consistence for training and constructing the classifier. Meta heuristic bee’s algorithm will be used as Feature Selection technique with the support of two of statistical ranking filters. The ranking of features with bee give an optimized ordering to the most critical and intrinsic features in the space of training and constructing classifier. The results obtained by constructing the most reliable classifiers Interactive Dichotomizer 3 classifier (ID3), Naïve Bayesian Classifier (NB), Artificial Neural Network (ANN) and Support Vector Machine (SVM) depending on both updated dataset and bee’s ranked features sets give effective efficiency in reducing false alarms and increasing detection rates.

Published
2013-06-30
How to Cite
[1]
Soukaena H. Hashem, Shatha Habeeb, and Besmah M. Khalil, “Enhance Network Intrusion Detection System Using Bee Algorithm ”, JMAUC, vol. 5, no. 1, pp. 108-116, Jun. 2013.
Section
Articles