Evaluation the efficiency of machine learning boosting methods for estimating the water quality index of the Zayandeh Rood River

Articles in Press

Document Type : Research Paper

Authors

Department of Water Science and Engineering. College of Agriculture, Isfahan University of Technology, Isfahan, Iran

Abstract

Regarding climate change, global warming, and the reduction of water resources, surface water quality is of great interest to river engineers as surface water is one of the most important water resources in the world. Since the most widely used water quality index is the WQI index, the goal and importance of this research are to model the WQI using two machine learning boosting methods in the Zayandeh Rood River, Gradient Boosting and XGBoost. First, based on water quality data, the water quality index (NSFWQI) was calculated, and then, for modeling, input data including water quality characteristics of 8 stations over 31 years and the calculated WQI were used. In this study, the model was coded in the Google Colab environment, and 80% of the data was used in the training phase and the remaining 20% in the evaluation phase. Based on the results of the evaluation criteria of coefficient of determination (R2), mean absolute error (MAE), maximum error (ME), mean square error (MSE), root mean square error (RMSE), and normalized root mean square error (NRMSE), the optimal model was selected. The results of the study showed that in all stations except one station among the models used, the GB performed better than the XGBoost according to the model evaluation criteria. The results also showed that to save time and cost, and also to optimally manage water quality characteristics, the selection of the number 3 series, in which three characteristics are used to estimate the WQI, was the best combination.

Keywords

Main Subjects

Iranian Journal of Soil and Water Research

Articles in Press, Accepted Manuscript
Available Online from 12 May 2025

Files

Share

How to cite

Statistics