Ließ, M.; Hitziger, M. & Huwe, B. (2014): <b>The Sloping Mire Soil-Landscape of Southern Ecuador: Influence of Predictor Resolution and Model Tuning on Random Forest Predictions</b>. <i>Applied and Environmental Soil Science</i> <b>2014</b>(603132), 10 pages.
Resource Description
Title:
The Sloping Mire Soil-Landscape of Southern Ecuador: Influence of Predictor Resolution and Model Tuning on Random Forest Predictions
email:
mareike.liess <at> ufz.de
Helmholtz Centre for Environmental Research - UFZ
Department of Soil Physics
D-06120 Halle (Saale), Theodor-Lieser-Str. 4
Germany
Individual:
Martin Hitziger
Contact:
email:
webmaster <at> tropicalmountainforest.org
Individual:
Bernd Huwe
Contact:
email:
bernd.huwe <at> uni-bayreuth.de
University of Bayreuth
Dept. of Geosciences/ Soil Physics Devision
95440 Bayreuth
Germany
Abstract:
The sloping mire landscape of the investigation area, in the southern Andes of Ecuador, is dominated by stagnic soils with thick organic layers. The recursive partitioning algorithm Random Forest was used to predict the spatial water stagnation pattern and the thickness of the organic layer from terrain attributes. Terrain smoothing from 10 to 30m raster resolution was applied in order to obtain the best possible model. For the same purpose, several model tuning parameters were tested and a prepredictor selection with the R-package Boruta was applied. Model versions were evaluated and compared by 100 repetitions of the calculation of the residual mean square error of a five-fold cross-validation. Position specific density functions of the predicted soil parameters were then used to display prediction uncertainty. Prepredictor selection and tuning of the Random Forest algorithm in some cases resulted in an improved model performance.We therefore recommend testing prepredictor selection and tuning to make sure that<br/>
the best possible model is chosen.This needs particular emphasis in complex tropical mountain soil-landscapes which provide a real challenge to any soil mapping approach but where Random Forest has proven to be successful due to the testing of model tuning and prepredictor selection.
Additional Infos:
How to cite this article:<br/>
Mareike Ließ, Martin Hitziger, and Bernd Huwe, “The Sloping Mire Soil-Landscape of Southern Ecuador: Influence of Predictor Resolution and Model Tuning on Random Forest Predictions,” Applied and Environmental Soil Science, vol. 2014, Article ID 603132, 10 pages, 2014. doi:10.1155/2014/603132
email:
mareike.liess <at> ufz.de
Helmholtz Centre for Environmental Research - UFZ
Department of Soil Physics
D-06120 Halle (Saale), Theodor-Lieser-Str. 4
Germany
