International Research Journal of Education and Technology

Increase in networking and communication technologies has resulted in improved lifestyle of people. However, the large and sensitive information being transmitted online has become a target for fraudsters. This has resulted in the need for an effective intrusion detection system to safeguard the critical information. This work presents a two phased intrusion detection model, SIBC, that aims to automatically handle data imbalance and also provide effective intrusion detection. Experiments were performed with KDD cup data, NSL-KDD data, and UNSW-NB15 data. Comparisons indicate that the SIBC model performs effective detection of intrusions with accuracy levels with accuracy levels greater than 90% indicating highly effective predictions.

Keywords: Network Intrusion Detection; Ensemble Modeling; Boosting; Sampling; Data Imbalance