Keywords: saturated soil hydraulic conductivity, statistical regressions, prediction, genetic algorithm, support vector machines.

Saturated hydraulic conductivity is one of the key parameters in soil physics and hydrological modeling. The objective of the study, presented in this paper is to explore the applicability of a support vector machine approach with different kernel functions for predicting the field saturated soil hydraulic conductivity of sandy soil and to compare this approach with the nonlinear statistical regression approach based on basic saline and alkaline soil data sets. Considering the significance of soil properties, both methods used the following classes of input soil data, which are easily measurable in the laboratory: hydraulic conductivity, clay to silt ratio, liquid limit, hydrocarbonate anions, chloride ions, and calcium carbonate content. A genetic algorithm is used to determine optimal values of the free support vector machine parameters for different kernel functions. Compared with the regression results and an associated selection of soil type, the excellent performance of support vector machine with a radial-basis-kernel-based model demonstrated the potential to function as a useful tool for the indirect estimation of field saturated soil hydraulic conductivity to assess the maximum obtainable prediction accuracy.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £45 +P&P)