Time Series Traffic Flow Prediction with Hyper-Parameter Optimized ARIMA Models for Intelligent Transportation System

Abstract

Intelligent Transportation System (ITS) has become the need of the day to manage heavy traffic problems due to the exponential growth of road transportation. This is also very much essential for building the smart cities and to improve the comfort of the vehicle drivers. The electric and autonomous vehicles are going to be the future transport systems for which we need an intelligent traffic management system. This requires a lot of growth in infrastructure. The integration of technologies such as Sensors, Internet of Things (IoT), Cloud Computing, etc. has to be done for this. The traffic prediction is one of thekey requirement for establishing the ITS. In this paper we present our study on ARIMA model with optimized hyper-parameter using grid search technique for traffic flow predictions. The model validation is done on the whole day traffic flow, morning and evening peak time traffic flow datasets. The prediction results show good performance metrics with RMSE of 8.953, 11.007 and 11.837 for those three datasets.