Feature Selection Using Modified Charged System Search for Intrusion Detection Systems in Cloud Environment
International Journal of Control Theory and Applications, 9(41), 1012-1021
Co-Authors: Partha Ghosh, Rupesh Kumar, Santanu Phadikar
Abstract
Cloud Computing is one of the most disruptive technology of modern age. It has revolutionized the concept of scalable services offered to the users over Internet. As a result of its extensive use, the vulnerabilities present in Cloud owes a threat to its security. Thus an Intrusion Detection System (IDS) is necessary in order to efficiently detect intrusions and protect the Cloud. The amount of incoming traffic is generally huge and it becomes necessary to reduce its volume by selecting only the relevant features for detecting intrusions. In this paper the authors have proposed a Modified-Charged-System-Search (MCSS) algorithm for feature selection. The proposed algorithm effectively manages the exploration-exploitation tradeoff and is fast in convergence. The selected features improve accuracy of the IDS and makes it more reliable. The proposed model is evaluated on KDD Cup 99 and NSL-KDD dataset. The results prove that MCSS algorithm is successful in selecting optimal feature subset for proposed IDS.