Interpretable machine learning framework for domestic wastewater management toward SDG 6.3
Issued Date
2026-06-01
Resource Type
eISSN
25901230
Scopus ID
2-s2.0-105034992351
Journal Title
Results in Engineering
Volume
30
Rights Holder(s)
SCOPUS
Bibliographic Citation
Results in Engineering Vol.30 (2026)
Suggested Citation
Taweesan A., Kanabkaew T., Eamrat R., Pussayanavin T., Surinkul N., Sukthanapirat R., Polprasert C. Interpretable machine learning framework for domestic wastewater management toward SDG 6.3. Results in Engineering Vol.30 (2026). doi:10.1016/j.rineng.2026.110375 Retrieved from: https://repository.li.mahidol.ac.th/handle/123456789/116161
Title
Interpretable machine learning framework for domestic wastewater management toward SDG 6.3
Corresponding Author(s)
Other Contributor(s)
Abstract
Safely managed domestic wastewater is a fundamental component of urban environmental health and a core target of Sustainable Development Goal (SDG) 6.3. This study hypothesizes that domestic wastewater management safety is governed by the interaction of operational capacity, institutional performance, and spatial system characteristics. This study examines the heterogeneity of domestic wastewater management systems and identifies the operational, institutional, and spatial determinants distinguishing safely managed from unsafely managed systems. An interpretable, machine learning–based analytical framework was developed and applied to a national dataset covering 200 cities using standardized indicators of operational capacity and institutional and service performance aggregated into composite Z-score indices. Correlation analysis and K-means clustering were used to characterize system heterogeneity, followed by a J48 decision-tree classifier to derive interpretable decision rules. The results reveal pronounced heterogeneity across cities and identify a limited set of model-derived, conditional operational and institutional conditions associated with transitions between safe and unsafe domestic wastewater management. Inadequate wastewater collection performance, constrained human-resource capacity, and spatially extensive service areas under limited institutional capacity emerge as dominant risk conditions for unsafe management. The decision-tree model provides directly interpretable rules that support management prioritization. This study provides an interpretable analytical framework for assessing domestic wastewater management performance and identifying determinants of safe management. The study also provides evidence-based insights that enable wastewater utilities, municipal authorities, and policy makers to classify cities according to management risk level, prioritize targeted operational and infrastructure interventions, and support progress toward safely managed sanitation services under SDG 6.3.