Document Type : Research Paper
Authors
1
Ardakan uni
2
Associate Professor, Department of Horticulture, Faculty of Agriculture and Natural Resources, Ardakan University, Ardakan, Iran.
3
Associate Professor, Department of Nature Engineering, Faculty of Agriculture & Natural Resources, Ardakan University, P.O. Box 184, Ardakan, Iran.
Abstract
Irrigation water assessment using conventional methods is usually costly and time-consuming, as it requires the measurement of several parameters. Therefore, the development of accurate and reliable models can help reduce costs and optimize water resource management in agriculture. The parameters Sodium Adsorption Ratio (SAR), Total Hardness (TH), and Residual Sodium Carbonate (RSC) play a vital role in agricultural activities, but their measurement requires specialized equipment and considerable time. The aim of this study is to predict SAR, TH, and RSC values based on measured Electrical Conductivity (EC) and pH using different machine learning algorithms. For this purpose, seven algorithms—including Decision Tree, K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Cubist, Random Forest, Super Learning, and XGBoost—were examined. In this study, data from 132 water samples with laboratory-measured EC, pH, SAR, RSC, and TH values were used. The results showed that the use of Super Learning significantly improved prediction accuracy compared to other algorithms. Specifically, the R² values obtained using both EC and pH simultaneously for predicting TH, RSC, and SAR were 0.91, 0.88, and 0.68, respectively, which were higher than those of other methods. Using EC alone yielded R² values of 0.87 for TH, 0.86 for RSC, and 0.73 for SAR, while using pH alone reduced the R² values to 0.12, 0.11, and 0.09, respectively. The findings of the present study indicate that with simple EC and pH measurements combined with machine learning algorithms—especially Super Learning—it is possible to accurately predict TH, RSC, and SAR parameters, thereby reducing costs and enabling faster decision-making in water resource management.
Keywords
)