Title: Monthly stream flow prediction: the power of ensemble machine learning-based decision support models

Authors: Hamit Erdal; Ersin Namli

Addresses: Faculty of Security Sciences, Gendarmerie and Coast Guard Academy, Ahlatlıbel, 06805 Çankaya/Ankara, Turkey ' Department of Industrial Engineering, Faculty of Engineering, Istanbul University-Cerrahpasa, Istanbul, Turkey

Abstract: Predicting stream flow is a vital milestone in planning and managing water resources, and is important to researchers and hydrologists. Unpredicted stream flow threads cultivated areas, dams and riverside lands. Recently, the increasing popularity of machine learning (ML) methods including ensemble methods in hydrological prediction is noticeable. In this study, five single and three ensembles ML-based 18 prediction models and six performance evaluation measurements are utilised for monthly stream flow prediction. It proved that models developed by stacking and voting ensemble ML methods have higher prediction accuracy. As a conclusion, this paper has presented the promising endeavour of incorporating sentiment regression into stream flow prediction.

Keywords: stream flow prediction; water resources; artificial intelligence; machine learning; ensemble learning; Turkey.

DOI: 10.1504/IJHST.2023.131834

International Journal of Hydrology Science and Technology, 2023 Vol.16 No.1, pp.17 - 36

Received: 04 May 2021
Accepted: 25 Jan 2022
Published online: 04 Jul 2023 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article