Prediction of transportation systems energy demand using intelligent techniques

Energy preservation and reduction of energy consumption is a major challenge faced by countries globally. By utilizing different Machine Learning (ML) models such as Stacking Ensemble Machine Learning (SEML) and Artificial Neural Network (ANN) algorithms, this paper aims to develop machine learning models to predict Transportation Energy Demand (TED) for different countries. Another major objective is to identify the most significant features that could help in predicting such information. Unlike the previous works that focused on specific countries, the proposed framework is generally applicable to many counties. Many experiments have been applied using benchmark dataset obtained from the Organization for Economic Co-operation and Development (OECD) in order to predict the TED for fourteen countries using available historical data from 2000 to 2020. The results showed that the SEML model performed better with an R2 of 0.9963 using the most significant feature list. Moreover, based on projections for 2040 data from recent research in Turkey, our study presents future predictions for Turkey’s TED using our generalized framework. We estimate that TED will reach approximately 41 Million Tons of Oil Equivalent (MTOE) by 2040.