Application of novel binary optimized machine learning models for monthly streamflow prediction

Abstract Accurate measurements of available water resources play a key role in achieving sustainable environment society. Precise river flow estimation is an essential task for optimal use hydropower generation, flood forecasting, and best utilization engineering. The current paper presents the development verification prediction abilities new hybrid extreme learning machine (ELM)-based models coupling with metaheuristic methods, e.g., Particle swarm optimization (PSO), Mayfly algorithm (MOA), Grey wolf (GWO), simulated annealing (SA) monthly streamflow prediction. Prediction precision standalone ELM model was compared two-phase optimized state-of-the-arts models, ELM–PSO, ELM–MOA, ELM–PSOGWO, ELM–SAMOA, respectively. Hydro-meteorological data acquired from Gorai Padma Hardinge Bridge stations at River Basin, northwestern Bangladesh, were utilized as inputs this study to employ form seven different input combinations. model’s performances are appraised using Nash–Sutcliffe efficiency, root-mean-square-error (RMSE), mean absolute error, percentage error determination coefficient. tested results both reported that ELM–SAMOA ELM–PSOGWO offered accuracy streamflows models. Based on local data, reduced RMSE ELM, by 31%, 27%, 19%, 14% station 29%, bridge station, testing stage, In contrast, based external improves 20%, 5.1%, 6.2%, 4.6% confirmed superiority over single model. overall suggest can be successfully applied modeling either or hydro-meteorological datasets.