Utilizing Neural Networks For Effective Intrusion Detection

Computer security, and intrusion detection in particular, has become increasingly important in today's business environment, to help ensure safe and trusted commerce between business partners as well as effective organisational functioning. Various approaches to intrusion detection are currently being utilized, but unfortunately in practice these approaches are relatively ineffective. New ways and means must, therefore, continuously be researched and defined. This paper will propose a proactive and dynamic model, based on trend analysis, fuzzy logic and neural networks that could be utilized to minimise and control intrusion to an organisation's computer system. The model will be based on the assumption that each user is unique and leaves a unique footprint on a computer system when using it. A back-propagation neural network was trained to implement this idea. KEY WORDS Computer security; intrusion detection; intrusion detection systems; fuzzy logic; neural network; pattern...

• Type: Conference
• Year: 2004
• Where: ISSA
• Authors: Martin Botha
• Keywords: Computer Security, Intrusion Detection, ISSA 2004, ISSA 2008, Neural Network