Comparison of Artificial Neural Network Methods and Support Vector Machine in Predicting Water Quality Parameters of Dinachal River, IRAN

Articles in Press

Document Type : Research Paper

Authors

1 Irrigation and Reclamation Engineering Department, Faculty of Agriculture, University of Tehran, Karaj, Iran.

2 Professor, Department of Renewable Energies and Sustainable Resources Engineering, University of Tehran, Tehran, Iran.

Abstract

Predicting water quality parameters plays a crucial for better monitoring of ecosystems of rivers and their sustainability. Alongside this، conventional prediction models are not able to capture the non-linearity and non-stationary inherence of water quality datasets. In recent years، the rapid development of machine learning methods has transformed the water quality prediction fields. In this study، water quality parameters for the Dinachal River in Guilan province have been assessed and predicted. Two models based on Artificial Neural Networks (ANN) and Support Vector Regression (SVR) were utilized to predict nine water quality parameters as TDS، EC، pH، Cl، SO4، HCO3، Ca، Mg، and Na with monthly timesteps between 2006 and 2018. Then، the model’s performance was evaluated using RMSE، MSE، and MAPE indices. According to the results، the SVR model was superior in predicting TDS and Mg parameters with an RMSE Index of 2.03 and 0.062، respectively. Simultaneously، ANN had a slightly better accuracy in the prediction of remaining parameters. However، prediction results for both models in the case study were satisfactory. In addition، the SVR model predicted TDS with a MAPE of 0.007، which was the best compared to other parameters. At the same time، the ANN model had better performance in predicting EC with a MAPE of 0.001. Prediction results for Cl had also the lowest accuracy among water quality parameters with an RMSE of 0.055 and 0.052 for SVR and ANN، respectively. Methods utilized in this study can be effective in predicting water quality parameters of Dinachal river.

Keywords

Main Subjects

Iranian Journal of Soil and Water Research

Articles in Press, Accepted Manuscript
Available Online from 05 February 2024

Files

Share

How to cite

Statistics