Volume 13, Issue 25 (5-2022)
J Watershed Manage Res 2022, 13(25): 74-85 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
bizhan I, piri H, tabatabaii M, piri J. (2022). Comparison and Application of Artificial Neural Network, Support Vector Machine and Decision Trees in Predicting the Hydraulic Conductivity of Soil Saturation (Case Study: Hirmand City). J Watershed Manage Res. 13(25), 74-85. doi:10.52547/jwmr.13.25.74
URL: http://jwmr.sanru.ac.ir/article-1-1111-en.html
URL: http://jwmr.sanru.ac.ir/article-1-1111-en.html
1- zabol university
Abstract: (1645 Views)
Extended Abstract
Introduction and Objective: Direct measurement of soil hydraulic conductivity is time consuming and costly, and sometimes the results are unreliable due to trial and error. This parameter can be estimated using early soil parameters. The present study was conducted to predict the hydraulic conductivity of soil saturation using decision tree methods, support vector machine and artificial neural network in Helmand city.
Material and Methods: For this purpose, 130 soil samples were collected from the surface (0-30 cm) and transferred to the laboratory for testing and analysis. In the laboratory, the parameters of hydraulic conductivity of soil saturation, percentage of stress, sand and silt, organic matter, acidity, electrical conductivity and calcium carbonate were measured. It was then estimated using measurement parameters and using decision tree models, artificial neural network, and saturated hydraulic guidance support vector machine. In order to evaluate the models, the criteria of explanatory coefficient, square mean error and absolute mean error were used.
Results: The results showed that the decision tree model with the highest coefficient of explanation (0.83) and the lowest value of the mean square of error and absolute error of average (0.0026 and 0.0019) is the best model for predicting the hydraulic conductivity of soil saturation in Hirmand region. Also, the results of data sensitivity analysis using artificial neural network model showed that sand percentage, lime percentage, silicate percentage and acidity are the most important factors affecting the hydraulic conductivity of soil saturation in Hirmand city, respectively.
Conclusion: The results show the very good performance of artificial intelligence methods in predicting hydraulic conductivity of soil saturation. In cases where it is not possible to measure hydraulic conductivity, hydraulic conductivity can be estimated using early soil parameters and artificial intelligence methods.
Introduction and Objective: Direct measurement of soil hydraulic conductivity is time consuming and costly, and sometimes the results are unreliable due to trial and error. This parameter can be estimated using early soil parameters. The present study was conducted to predict the hydraulic conductivity of soil saturation using decision tree methods, support vector machine and artificial neural network in Helmand city.
Material and Methods: For this purpose, 130 soil samples were collected from the surface (0-30 cm) and transferred to the laboratory for testing and analysis. In the laboratory, the parameters of hydraulic conductivity of soil saturation, percentage of stress, sand and silt, organic matter, acidity, electrical conductivity and calcium carbonate were measured. It was then estimated using measurement parameters and using decision tree models, artificial neural network, and saturated hydraulic guidance support vector machine. In order to evaluate the models, the criteria of explanatory coefficient, square mean error and absolute mean error were used.
Results: The results showed that the decision tree model with the highest coefficient of explanation (0.83) and the lowest value of the mean square of error and absolute error of average (0.0026 and 0.0019) is the best model for predicting the hydraulic conductivity of soil saturation in Hirmand region. Also, the results of data sensitivity analysis using artificial neural network model showed that sand percentage, lime percentage, silicate percentage and acidity are the most important factors affecting the hydraulic conductivity of soil saturation in Hirmand city, respectively.
Conclusion: The results show the very good performance of artificial intelligence methods in predicting hydraulic conductivity of soil saturation. In cases where it is not possible to measure hydraulic conductivity, hydraulic conductivity can be estimated using early soil parameters and artificial intelligence methods.
Type of Study: Research |
Subject:
مديريت حوزه های آبخيز
Received: 2020/09/23 | Accepted: 2020/12/15
Received: 2020/09/23 | Accepted: 2020/12/15
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
