Machine Learning Applications in Industrial Pollution Monitoring and Wastewater Management: A Bibliometric Analysis

Abstract

Introduction: Industrial pollution and wastewater contamination represent escalating global environmental crises, yet a systematic, evidence-based understanding of how machine learning (ML) is being applied to address these challenges remains lacking. This study maps the intellectual landscape and emerging research frontiers in ML-based environmental monitoring and wastewater management through a bibliometric lens. Materials and Methods: A total of 171 documents retrieved from the Scopus database (2008–2026) were analyzed using bibliometric methods. VOSviewer was employed to construct co-occurrence keyword networks, overlay maps, and density visualizations. Quantitative analysis of publication trends, document types, top journals, and citation impact. Results: Publication output grew exponentially from 2 articles in 2008 to 58 articles in 2025, with 2026 already yielding 34 records as of May. Four thematic clusters were identified: (1) wastewater treatment process optimization using artificial neural networks and deep learning; (2) water quality and pollutant detection; (3) industrial wastewater and environmental monitoring; and (4) emerging contaminants and greenhouse gas modeling. XGBoost, explainable artificial intelligence (XAI), and data-driven modeling represent the most rapidly growing thematic frontiers. Conclusion: The field is transitioning from predictive accuracy toward interpretability, real-time deployment, and multi-hazard environmental risk assessment. Critical research gaps include standardized benchmark datasets, cross-domain model transferability, and equitable global collaboration structures. This study provides an evidence-based roadmap for future ML-integrated environmental research.