Green Clustering Analyzing Logistics Performance and Carbon Emissions with K -Means and Gaussian Mixture Models
Jiratchaya Chomjinda, Janjira Piladaeng, Tanayot Kulthon, Pattharaporn Thongnim · 2024
This study employs K-Means and Gaussian Mixture Model (GMM) to analyze the relationship between logistics performance index (LPI) and carbon emissions across various regions from 2007 to 2018. Using data on LPI and CO2emissions, it groups Asian countries into two clusters to uncover patterns that reveal the relationship between logistics efficiency and environmental sustainability. K-Means clustering, optimized through the Elbow method and Silhouette Score, highlights distinct groupings based on logistics performance, while GMM, chosen for their probabilistic approach and optimized by the Bayesian Information Criterion (BIC), offer insights into the complex nature of data distribution. Among evaluation metrics, the Silhouette Score, which is associated with K-Means, stands out as the best for validating the clustering results. This offers a foundation for understanding the environmental footprint of the logistics section. In addition, the research represents the importance of integrating health considerations into logistics and environmental strategies, recommended for approaches that support environmental factor, logistic system, and public health.