A Hybrid Intrusion Detection System Based on C5.0 Decision Tree Algorithm and One- Class SVM with CFA

Cyber security threats have become increasingly sophisticated and complex. Intrusion detection which is one of the major problems in computer security has the main goal to detect infrequent access or attacks and to protect internal networks. A new hybrid intrusion detection method combining multiple classifiers for classifying anomalous and normal activities in the computer network is presented. The misuse detection model is built based on the C5.0 Decision tree algorithm and using the information collected anomaly detection model is built which is implemented by one class Support Vector Machine (SVM).The key idea is to take advantage of cuttlefish algorithm (CFA).In the proposed algorithm, Cuttlefish can find best selected features to remove the redundant and irrelevant features to evaluate the accuracy of classification. Integration of multiple algorithms helps to get better performance. The Experimental results are performed on NSL-KDD Dataset, and it is shown that overall performance of the proposed approach is improved in terms of detection rate and low false alarms rate in comparison to the existing techniques.