Forecasting Transport Emissions through Machine Learning Approaches: Insights for Sustainable Logistics
Forecasting Transport Emissions through Machine Learning Approaches: Insights for Sustainable Logistics

Summary/Abstract: This study examines the environmental impacts of the logistics sector operating under growing sustainability pressures driven by increasing environmental awareness and sought to analyse the interactions of economic factors that determine transport-related emissions. The originality of the research lies in the comparative application of Machine Learning (ML) approaches – Random Forest and Gradient Boosting – alongside conventional Linear Regression models to predict and explain transport emissions. The methodology employed World Bank data covering the period 2000-2023 for 81 countries in the high-income and upper-middle-income groups. In the analysis, Transport Emissions served as the dependent variable, while Exports of Goods and Services as a Percentage of GDP, GDP Growth Rate, and GDP per Capita were employed as independent variables. Descriptive statistics revealed profound disparities across the examined countries, both in terms of economic indicators and transport emissions. The forecasting results indicated that the Random Forest and Gradient Boosting models provided the most accurate predictions, showing noticeably greater explanatory power than Linear Regression. These findings provided evidence that, in the formulation of policies aimed at reducing logistics-related emissions, the complex and nonlinear interplay between economic growth and global trade can be more accurately modelled and predicted using advanced machine learning algorithms rather than traditional econometric methods.

• Details
• Contents