IMPROVED GENETICALLY OPTIMIZED NEURAL NETWORK ALGORITHM FOR CLASSIFICATION OF DISTRIBUTED DENIAL OF SERVICE ATTACKS

Abstract

Distributed Denial of Service (DDoS) attack has continued to grow dynamically and has increased significantly to date. This form of attack is usually carried out by draining the available resources in the network as well as flooding the package with a significant intensity so that the system becomes overloaded and stops. This research proposes a classification of DDoS attack using neural network-based genetic algorithm (NNGA). The genetic algorithm was used to optimize neural network for the detection of DDoS attacks in order to improve the effectiveness and efficiency of classification accuracy and performance. To improve the NNGA, a fitness function was introduced in genetic algorithm that improved the performance of NNGA. The features of DDoS attacks from KDD 99 intrusion detection datasets were obtained to train the NNGA. The results obtained from the study indicate that the technique performed optimally in DDoS attack 2 recording the following; 98.58% and 0.351 respectively for accuracy and false positive rate. Therefore, revealed that the enhanced genetically optimized neural network algorithm has better accuracy and lower false positive rate in comparison with the conventional neural networks.