Publication: Forecasting multi-step-ahead reservoir monthly and daily inflow using machine learning models based on different scenarios
Date
2023
Authors
Ibrahim K.S.M.H.
Huang Y.F.
Ahmed A.N.
Koo C.H.
El-Shafie A.
Journal Title
Journal ISSN
Volume Title
Publisher
Springer
Abstract
Dam reservoir operations are a critical issue for decision-makers in maximizing the use of water resources. Artificial Intelligence and Machine Learning models (AI & ML) approaches are increasingly popular for reservoir inflow predictions. In this study, the multilayer perceptron neural network (MLP), Support Vector Regression (SVR), Adaptive Neuro-Fuzzy Inference System (ANFIS), and the Extreme Gradient Boosting (XG-Boost), were adopted to forecast reservoir inflows for the monthly and daily timeframes. Results showed that: (1) For the monthly timeframe, all the four models were proficient in obtaining efficient monthly reservoir inflows by scoring at least an R� of 0.5
with the XG-Boost ranked as the best model, followed by the MLPNN, SVR, and lastly ANFIS. (2) the XG-Boost still outperforms all other models for forecasting daily inflow
but however, with reduced performance. The models were still ranked in the same order, with the ANFIS showing very poor performance in scenario-2, scenario-3, and scenario-4. (3) For daily inflows, the best scenarios are scenario-5, scenario-6, scenario-7 as the models were trained based on the 1,3,5, days-lag forecasted inflow, and overall, the XG-Boost outperforms all the other models. � 2022, The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.
with the XG-Boost ranked as the best model, followed by the MLPNN, SVR, and lastly ANFIS. (2) the XG-Boost still outperforms all other models for forecasting daily inflow
but however, with reduced performance. The models were still ranked in the same order, with the ANFIS showing very poor performance in scenario-2, scenario-3, and scenario-4. (3) For daily inflows, the best scenarios are scenario-5, scenario-6, scenario-7 as the models were trained based on the 1,3,5, days-lag forecasted inflow, and overall, the XG-Boost outperforms all the other models. � 2022, The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.
Description
Keywords
Adaptive neuro-fuzzy inference system (ANFIS) , Extreme Gradient Boosting (XG-Boost) , Grid Search optimizer , Hyper-parameters , Inflow Forecast , Machine learning , Multilayer Perceptron neural network (MLPNN) , Support Vector Regression (SVR) , Adaptive boosting , Decision making , Forecasting , Fuzzy inference , Fuzzy neural networks , Fuzzy systems , Multilayer neural networks , Multilayers , Reservoirs (water) , Water resources , Adaptive neuro-fuzzy inference , Adaptive neuro-fuzzy inference system , Extreme gradient boosting (XG-boost) , Gradient boosting , Grid search , Grid search optimizer , Hyper-parameter , Inflow forecast , Machine-learning , Multilayer perceptron neural network , Multilayers perceptrons , Neuro-fuzzy inference systems , Perceptron neural networks , Search optimizer , Support vector regression , Support vector regressions , Machine learning