An Intrusion Detection System Architecture Based on Neural Networks and Genetic Algorithms

The Computer Security is a part of Computer Science that addresses all the problems that exists within the computer security domain. Now days, more than ever, the computer viruses, worms, hackers, crackers, electronic eavesdropping and electronic fraud, have become the new problems that the Network Computer Security Science must solve. The Intrusion Detection System has become a needful component in a well formed network security policy. Although, intrusion detection is not a new technology, since it has been research subject for almost 30 years now, it has been liable to the false positive problem i.e. base-rate fallacy. Furthermore, the lack to detect a novel attack is probably a more serious shortcoming of intrusion detection. To address those shortcomings, intrusion detection science has to implement more complex techniques from Artificial Intelligence field as Artificial Neural Networks into the system design. However AI is not enough and research has involved a Genetic Algorithm as a method to optimize an artificial neural network. The preliminary research in this thesis has been carried out by literature survey and observation. Major goals were to describe and design an advisory intrusion detection system model that has focus on shortcomings that were introduced above. Additionally, general knowledge of Intrusion Detection System, Neural Networks and Genetic Algorithm were studied. The thesis gives an understandable overview of mentioned technologies and methods. The result of the research is the model called “Efficient Artificial Neural Network Genetic Algorithm Intrusion Detection System”