Development and application of modeling techniques to estimate the unsaturated hydraulic conductivity

The knowledge of the unsaturated hydraulic conductivity ( K ) is very essential for the various fields of water resources, irrigation, and hydrology. It is also important to know the phenomena of water movement on the ground constantly. In order to better forecast unsaturated hydraulic conductivity, this article reports the comparison of efficacy of five distinct soft computing approaches: support vector machine (SVM), random forest, Gaussian process (GP), gene expression techniques, and multivariate adaptive regression spline. Three kernels function (Poly, RBF, and PUK) were used in SVM and GP modeling techniques. For fulfill this aim, experimentation has been performed using mini-disc infiltrometer in 20 locations in Ghaggar basin. Total 240 observations were collected, and out of which, 70% were used for training the model and remaining for testing. The input variables of this investigation were sand, clay, silt, bulk density ( ρ ) and moisture content and output variable was K . The result of modeling techniques suggests that PUK kernel with SVM was superior to the other modeling techniques. This implies that these computational methods can be used to make estimates about the values of K at any time.