Determination of factors affecting logistics performance index with machine learning methods

The increasing complexity of global trade and the internationalization of supply chains have rendered countries' logistics performance a strategic priority. In this context, the logistics performance index (LPI), which enables a comparative assessment of national logistics capabilities, has become a crucial indicator for policymakers and researchers. The primary aim of this study is to identify the key determinants influencing the LPI and to evaluate their relative importance using machine learning algorithms. The LPI is employed as the dependent variable, while the independent variables include inflation rate, GDP growth, per capita GDP growth, human development index (HDI), carbon emissions, urbanization rate, population growth, trade volume of goods and services and transport services. The dataset was compiled from internationally recognized sources such as the World Bank and Our World in Data. During the feature selection and modeling phases, six different supervised learning algorithms were applied, and their prediction performance was comparatively evaluated. The results reveal that the most influential variables on the LPI are HDI, carbon emissions, population growth, transport services (export ratio) and carbon emissions from transportation. The findings contribute to policy-making processes aimed at improving countries' logistics performance.