A Comparative Study of Artificial Intelligence Algorithms for Network Traffic Prediction in VANET

Increasing the number of vehicles and their communications in smart cities is a critical issue that leads to road and network traffic. Traffic prediction with high accuracy and less complexity is a challenge in Intelligent Transportation System (ITS). In Artificial Intelligence (AI), Machine learning (ML) algorithms are promising solutions to prediction problems, and Deep Learning (DL) algorithms are used for more complicated issues. In this paper, we propose a comparative analysis of the prediction performance of the most five common AI algorithms used to solve classification problems. Different evaluation metrics are employed to analyze algorithms and get the most accurate one for selecting such problems. Simulation results on the Vehicular Ad-Hoc Network (VANET) dataset revealed that Random Forest (RF) as a traditional ML algorithm performed better than other algorithms in terms of accuracy (96%) and execution time (0.73 minute) for traffic prediction.