Unveiling sustainable solutions for municipal waste management through machine learning

Abstract

The increasing volume and complexity of municipal solid waste (MSW) in rapidly urbanizing regions pose significant environmental and public health challenges, especially in low- and middle-income countries. Despite efforts to improve municipal solid waste management (MSWM), many cities continue to rely on fragmented approaches that fail to ensure safe disposal and resource efficiency. This study addresses a critical research gap by adopting machine learning (ML) across the entire MSWM chain, including collection, transportation, treatment, and disposal, while aligning the outcomes with Sustainable Development Goal 11 (SDG11) on safely managed waste. Using data from 460 cities across nine Asian countries, the J48 decision tree algorithm was applied to classify MSWM practices as safe or unsafe. The model achieved 73 % training and 67 % validation accuracy, highlighting key determinants of safe management, including budget support, number of operators, and availability of collection vehicles. Findings reveal that 70 % of MSW in surveyed Thai cities is unsafely managed, with inadequate funding (below US$ 5 per ton) strongly linked to poor outcomes. This study contributes a transparent, interpretable ML model for data-driven decision-making and offers actionable insights for enhancing operational efficiency and advancing SDG11 targets through improved investment and resource allocation.