"Smart Urban Planning: An Intelligent Framework to Predict Traffic Using Stack Ensembling Approach "

The intelligent transportation system needs to accurately assess the volume of traffic in the environment in which it operates to ensure that people are moved in a timely and hassle-free manner. Forecasting systems allow drivers to identify the route that will take them to their destination with the slightest difficulty and the least time spent in congested regions. At present, both the corporate sector and government organizations require accurate and timely traffic flow information. There have been no significant efforts to enhance road traffic prediction by utilizing air pollution data. This paper aims to present a new method for predicting road traffic using data related to pollution. Our contribution to this research is twofold. Firstly, we compared ten regression approaches to determine which technique provides better results and accuracy. Secondly, we present a technique based on regression analysis approaches in which we choose those base learners who give better results on Level 1. These predictions are combined as an input to a Level 2 meta regressor. A method is proposed to show that it generates more satisfactory results than any of the regression procedures discussed previously. Compared with the various regression methodologies, the proposed method successfully lowers the mean square error, the relative absolute error, and the root mean square error and improves the R-squared value.