DefinePK

Abstract

Vehicular Adhoc Networks (VANETs) gain more attention in the modern world. The advancements in telecommunication technology has opened the door for deploying VANETs to make the roadway journey a more safer and comfortable one. The challenges such as mobility, network scalability and volatility make the VANETs more prone to be attacked by the intruders. In an open medium like VANETs, identifying the intruder is a difficult task. The behavior of intruders or malicious users is studied using various machine learning techniques. The work focuses on applying Support Vector Machine (SVM), a semi-supervised learning algorithm with Modified Fading Memory for predicting the behavior of the users of VANETs (SVM-MFM) and classifying the intruders from users of the network. This classification helps to isolate the intruders and the communication of the intruder with the VANET can be stopped, thus providing better resource utilization. This scheme is computationally fast in classifying the intruders with high ROCC of 98%.