Network Attack Detection Based on Combination of Neural, Immune and Neuro-fuzzy Classifiers
Abstract
Purpose: Imperfection of existing methods of intrusion detection and changing nature of malicious actions of the attackers lead computer systems to a compromised state. Therefore, it is important to identify new types of attacks and respond timely to them. Purpose: The development of a hybrid scheme of detection and classification of network attacks based on a combination of adaptive classifiers. Results: The generalized scheme of combining the classifiers to detect network attacks is offered. On its basis the software tool is developed which enables to analyze network traffic for anomalous network activity. To reduce the number of input features it is proposed to use the principal component analysis. The key features of the technique is a multi-level analysis of network traffic and using different adaptive modules while detecting the attacks. Computational experiments are performed on two public datasets using different means of combining classifiers. Practical relevance: Developed modules can be used for processing the data received from the security information and event management system.Published
2015-08-01
How to Cite
Branitskiy, A., & Kotenko, I. (2015). Network Attack Detection Based on Combination of Neural, Immune and Neuro-fuzzy Classifiers. Information and Control Systems, (4), 69-77. Retrieved from http://proceedings.spiiras.nw.ru/index.php/ius/article/view/4356
Issue
Section
Information security