A Systems-Based Explainable Machine Learning Framework for Environmental Sustainability

Environmental sustainability is a result of complex configurations of emissions, air pollution, and environmental infrastructure. This paper investigates environmental sustainability of 154 countries based on four environmental variables and the Natural Capital Protection (NCP) score. Principal component analysis (PCA), K-means clustering, prediction models, and explainable AI techniques, such as SHAP and LIME, in addition to network and persistent-homology analysis, are utilized as part of the methodology. Two principal components accounted for 74.70% of the total variance. Five environmental groups were identified to facilitate an interpretation of cross-country heterogeneity. PM2.5 turned out to be the most important model-assigned predictor of NCP, followed by methane and carbon emissions, while water access was the one with the lowest SHAP value on average. Network and persistent-homology analysis provided additional information about patterns of model attributions as well as about the relationships between environmental indicators.

Authors

Institutions

Publication Details

Journal
Systems
Published
2026-11-16
DOI
https://doi.org/10.3390/systems14111467
Primary Topic
Energy, Environment, Economic Growth
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

A Systems-Based Explainable Machine Learning Framework for Environmental Sustainability

Sercan Yavan, Sadullah Çelik
3 links
Systems
Energy, Environment, Economic Growth
article

A Systems-Based Explainable Machine Learning Framework for Environmental Sustainability

Sercan Yavan, Sadullah Çelik
article en

Abstract

Environmental sustainability is a result of complex configurations of emissions, air pollution, and environmental infrastructure. This paper investigates environmental sustainability of 154 countries based on four environmental variables and the Natural Capital Protection (NCP) score. Principal component analysis (PCA), K-means clustering, prediction models, and explainable AI techniques, such as SHAP and LIME, in addition to network and persistent-homology analysis, are utilized as part of the methodology. Two principal components accounted for 74.70% of the total variance. Five environmental groups were identified to facilitate an interpretation of cross-country heterogeneity. PM2.5 turned out to be the most important model-assigned predictor of NCP, followed by methane and carbon emissions, while water access was the one with the lowest SHAP value on average. Network and persistent-homology analysis provided additional information about patterns of model attributions as well as about the relationships between environmental indicators.

SystemsVol. 14(11)
Adnan Menderes University (TR)
Adnan Menderes Üniversitesi
Climate action
Energy, Environment, Economic Growth
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon