Advances in the applications of machine learning in environmental science and engineering

With the continuous advancement of environmental science and engineering, this field has increasingly entered the era of big data. As a pivotal branch of artificial intelligence, machine learning has been widely employed in environmental science and engineering owing to its superior capabilities in data processing and knowledge discovery. This review systematically examines the recent research progress and practical applications of machine learning in environmental science and engineering, encompassing diverse domains such as air pollution, water pollution, soil pollution, solid waste, noise pollution. The advantages and limitations of commonly used machine learning algorithms are comprehensively summarized. In addition, this paper discusses the existing challenges in current research and outlines prospects for future development. Among various algorithms, neural networks and random forests have gained significant traction due to their unique strengths and adaptability. Looking forward, the establishment of data and model sharing platforms, the development of ensemble learning approaches, and the integration of interdisciplinary technologies are identified as promising strategies to advance machine learning applications in environmental science and engineering.