A Hybrid Adaptive Development Algorithm and Machine Learning Based Method for Intrusion Detection and Prevention System

Abstract

Network Intrusion detection and prevention Systems (NIDPS) are employed in monitoring a network which safeguards user integrity, privacy thereby ensuring the data security and availability in a network. Such systems not only monitor the suspicious activities in a network but also used as control systems to eliminate the malicious users from the network. In this paper, a Hybrid Adaptive Development Algorithm and Machine Learning Algorithm (ADA-MLA) method is proposed to identify the malicious activities and eliminating them from the network. The deployment of honeypot-based intrusion is improved adaptive development algorithm. Machine learning algorithm has been employed in the Hybrid IDPS for learning the network data patterns which also identifies the maximum probable attacks in the network. The signatures for the DARPA 99 data set have been updated during the implementation of intrusion prevention system on a real-time basis. The hybrid method works on (i) classifying the attacks based on protocols and (ii) classifying the attacks on pre-determined threshold values. Hence, both known and unknown attacks can be easily captured in the proposed hybrid IDPS method which thereby achieves higher attack detection and prevention accuracy while compared to the conventional attack detection and prevention methodologies.