Network Intrusion Attack Detection and Prevention using Various Soft Computing and Deep Learning Techniques

Abstract

In modern times, there has been a substantial improvement in network malwares and threats offences, which constructs excellent attention of network secrecy and protection appearances. Due to the elevation of the technology, network wormhole attacks are converting immensely complicated. Such that the current exposure systems are not satisfactory adequate to discuss this detection and prevention. Therefore, the execution of a unique and efficient network invasion detection system would be important to resolving the problems. In this research, we utilize hybrid deep learning nbased methods, especially, Convolutional Neural Networks (CNN) as well as Recurrent Neural Networks (RNN) to produce a unique apprehension system which can distinguish different network intrusions. Additionally, we estimate the representation of the proposed approach that demonstrates different evaluation techniques, and we impersonate an association among the consequences of our recommended explication to find the best model for the specific network intrusion detection scheme. The experiment analysis of system has done with numerous machines learning algorithm that executed in Weka environment on KDDCUP99 and NSLKDD dataset. Finally we conclude system provides better detection than machine learning algorithms on well-known network datasets.