Network traffic classification using rough set theory and genetic algorithm

Network traffic classification by applications is very useful for many network activities which include Quality of service, security monitoring, abnormalities detection etc. In this paper, a hybrid approach for network classification is presented. The target is to classify the traffic flow data into different application categories. Rough set classification can help finding hidden patterns and identifying relationships that would not be found using statistical method. Genetic algorithm is applied to get the reducts. The approach is a supervised method based on training data. The result indicates that the approach can achieve good accuracy and trusty.