Development of local and global wastewater biochemical oxygen demand real-time prediction models using supervised machine learning algorithms

This study aims to shed light on a new understanding of global machine-learning (ML) prediction models for wastewater treatment plants (WWTPs). The paper evaluates the development local and predict influent biochemical oxygen demand (BOD5) in four WWTPs. proposes an integrated framework remote sensing ML techniques, specifically decision tree, random forest, adaptive boosting, gradient boost algorithms. modeling considered two cases models. model’s best score achieved 0.92 coefficient determination (R2), 6.56 mean absolute error (MAE), 1.00 R2, 0.08 MAE, respectively. In second case, 0.58 20.73 0.95 3.93 results showed that developed model reduced BOD5 test duration from five days only three hours. However, first failed achieve good predictions against other is due biological factors change characteristics one place another. concluded recommended be each WWTP separately. novelty this it investigates various testing performance different WWTPs than used train test. Also, technically discusses result complications prediction, which was never discussed or taken into consideration literature.