LAUSR

--------------------------------------------------------ABSTRACT----------------------------------------------------------Any violation of information security policy with malicious intent is regarded as an intrusion. The fast evolving new kind of intrusions poses a very serious threat to system security, although there has been the rapid development of several security tools to counter the growing threats, intrusive activities are still growing. Many Intrusion Detection models have been implemented since the concept of Intrusion Detection emerged, but the majority of the existing Intrusion detection models have many drawbacks which include but not limited to low accuracy in detection, high false alarm rates, adaptability weakness, inability to detect new intrusions etc. The main aim of this study is proposing a model that combined simple K-Means clustering Algorithms and Random Forest classification technique that will have minimum false alarms rate and high accuracy detection rate. The experiment was carried out in WEKA 3.8 using the NSL-KDD dataset to process the dataset and obtained the results. At the end of training and testing of the proposed study, the results indicated that the proposed approach achieved improved accuracy and reduced false alarm rates by 99.98% and 0.14% respectively.