Evaluating the Performance of Various Machine Learning Algorithms for Detecting DDOS Attacks in Vanets
The to be launched 5G networks can provide a very reliable, very fast communication in a vehicular ad hoc network (VANET). This fast network can provide various services such as safety, traffic management and driverless driving for drivers and commuters. But the characteristics of VANETs makes it prone to various types of security attacks which can cause loss of lives if not detected quickly. One of the most difficult type of attack to detect quickly is the Distributed Denial of Service. In these types of attacks multiple vehicles act as rogue nodes which carry out different types of attacks which are hard to tackle. An efficient detection algorithm is therefore needed for smooth and efficient running of VANET which can accurately detect the DDoS attacks. Machine Learning Algorithms are becoming very popular nowadays for their high accuracy rate in detecting DDoS attacks. Most of the machine learning model first train itself based on known dataset of attacks and then are used to predict known as well as unknown attacks based on its learned technique.This paper evaluates the performance of existing machine learning models for detecting DDoS attacks in VANETs. The main objective of this paper is to find an efficient classification model which can detect DDoS attacks in VANETs.
Keywords: DDoS attacks, VANETs, Random Forest, Decision Tree, SVM Using RBFDot, Linear Model, Neural Network, detection technique.