Y.B. Reddy (USA)
Intrusion Detection, Genetic Algorithm, GeneticOperators, fitness model
Hacking has increased a lot in the recent years to disturb, destroy, and access sensitive information. As malicious intrusions are a growing problem, we need a solution to detect the intrusions accurately. Network administrators are continuously looking for new ways to protect their resources from harm, both internally and externally. So there is a strong need for novel strategies for defense and new approaches to security. The present available techniques are useful to detect the known signatures (misuse detection). Many techniques were underway to detect the anomalies but had less success. Recent researchers [8-9, 12-16,28] were experimenting with probability theory, data mining techniques, fuzzy sets, rule-based systems, and neural networks. The false alarm rate was not decreased significantly. Genetic algorithms [29,30] were used in many areas and found satisfactory results. We conducted a literature search, which includes the past 10 years and found that very few were attempted to use genetic algorithms [1-7] for intrusion detection. So we attempted genetic algorithm technique to discover the consistent and useful patterns of programmer and system behavior and use the set of relevant system features to compute the classifiers that can recognize anomalies and known intrusions. We used the random generated data to test the behavior of intruders. The data set with axis attributes [14-16] were prepared for each individual (chromosome) and the fitness function defined for the problem. The simulations were presented as sample results.
Important Links:
Go Back
