Research advances in machine learning-aided accurate prediction of the impact of exogenous substances on anaerobic digestion.

Exogenous substances (including additives and pollutants), as key factors influencing anaerobic digestion (AD) systems, exert dual effects on microbial activity, community structure, and metabolic function. This review systematically evaluates the adaptability and limitations of existing mathematical models in predicting their complex mechanisms. On this basis, it emphasizes the modeling processes and applications of machine learning (ML) in AD, and outlines the progress and predictive performance of ML in research related to exogenous substances. The study summarizes the limitations of mathematical models in quantifying the impact of exogenous substances, provides a systematic review of analytical frameworks applying ML technology to the regulation of such substances in the field of AD, and clarifies both the potential and current constraints of ML in predicting exogenous substance effects. This work offers a theoretical reference for a deeper understanding of the mechanisms of exogenous substances and for constructing intelligent prediction models, while also providing methodological support for the optimization and regulation of AD processes.