Estimation of soil aggregate stability indices using artificial neural network and multiple linear regression models
Keyword(s):
During recent decades, an artificial intelligence system has been used for developing the pedotransfer functions (PTFs) for estimation of soil properties. In the present study, the capabilities of multiple linear regression (MLR) and artificial neural networks (ANNs) in developing PTFs for estimating mean weight diameter (MWD) from routine soil properties (P<sub>1</sub>) and combination of routine soil properties and fractal dimension of aggregates (P<sub>2</sub>) were evaluated. The results showed that the ANN model for estimating MWD is more accurate than the MLR model. Application of fractal dimension of aggregates as a predictor in both methods improved the accuracy of PTFs.
2017 ◽
Vol 10
(11)
◽
pp. 1-6
2017 ◽
Vol 44
(12)
◽
pp. 994-1004
◽
2018 ◽
Vol 117
◽
pp. 224-234
◽
2019 ◽
Vol 2019
◽
pp. 1-13
◽
2014 ◽
Vol 55
◽
pp. 29-36
◽
2021 ◽
Vol 18
(1)
◽
pp. 27