A Virtuous Reckoning on security enhancement of mobile Adhoc network using artificial intelligence

Abstract

Mobile ad hoc networks (MANET) have revolutionized our culture through their self-configured autonomous communications modes that have no framework and have thus been aimed at exploring means of making good use of the potentials of MANETs, more and better. The latest introduction of new machine learning technology has enabled the creation of the best protocols for artificial intelligence. Mobile ad hoc networks (MANETs) face different problems related to protection caused by malware attacks. Its transient existence makes nodes more sensitive to threats by nodes or by attackers as any node acts as a router to transfer data without centralized control. Therefore, to identify the wrong entry of misbehavior nodes, MANET requires very precise security policies. The networks perform better if the nodes are confident and cooperate correctly. This paper proposes an effective artificial intelligence algorithm-based security framework as an approach that recognizes and distinguishes insider threats in real-time by categorizing data packets as normal or abnormal, as well as identifying and detecting packet falling nodes through the vector support and logistic regression machine. The results suggest that LR was higher than SVM with an exact prediction rate of 97 %.LR is therefore more appropriate to identify malicious attacks in MANETs.