Integrating artificial intelligence for sustainable waste management: Insights from machine learning and deep learning

Abstract

As global waste production grows, sustainable waste management (WM) has become an issue for modern societies. This paper explores the integration of Artificial Intelligence (AI), particularly Machine Learning (ML) and Deep Learning (DL), to improve waste management (WM) systems by enhancing automation, classification accuracy, operational efficiency, and real-time decision-making. Current trends and potential future directions are identified with bibliometric and scientometric analysis, which assess methodologies and data in the field. By automating processes such as waste classification, sorting, and transportation, AI-driven models have the potential to optimize operational efficiency and reduce environmental impact. A comprehensive review of recent AI research in WM is presented, with a focus on their effectiveness, scalability, and limitations. Moreover, in the proposed framework, the data augmentation approach has been utilized to improve the model’s performance by increasing the amount of samples. Furthermore, the MobileNetV3 DL model is employed for feature extraction. Besides, the feature selection method − Harris Hawk Optimization (HHO) is also utilized to choose the best subset of features and reduce the irrelevant features. Then these selected features are fed into Machine Learning algorithms such as Decision Tree (DT), Logistic Regression (LR), and Random Forest (RF). In summary, this review highlights key case studies and research insights, offering a roadmap for future developments in AI-driven WM solutions.