Fig. 3

Genetic algorithm process for rule based dataset generation. Genetic algorithm is used to generate rule based dataset. Process works in four steps as initial population generation for features selected from CIC-IDS dataset. This dataset consist of more than 80 features out of that only 22 features are used to describe normal behavior of the packet in the network. Optimized set of features is finalized using fitness function. All the features with good fitness value are considered for rule dataset generation. Process continues until optimized set is found