Table of Contents
ISRN Soil Science
Volume 2012, Article ID 610567, 9 pages
http://dx.doi.org/10.5402/2012/610567
Research Article

An Ensemble of Neural Classifiers and Constructivist Algorithms in the Identification of Agricultural Suitability Complexes of Soils on the Basis of Physiographic Information

Faculty of Mining Surveying and Environmental Engineering, AGH University of Science and Technology, Al. Mickiewicza 30, 30-059 Cracow, Poland

Received 3 January 2012; Accepted 19 February 2012

Academic Editors: M. Cox, W. Ding, and Z. He

Copyright © 2012 Stanislaw Gruszczynski. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

The ensemble of classifiers for identification of agricultural suitability of soils on the basis of physiographic information was created in accordance with the stacking algorithm. It is comprised of five neural networks of various structures. The deciding element was a neural classifier optimised on the basis of input vectors composed of the indications of five classifiers making up the lower level. Among the architectures studied, the best result was achieved using the Radial Basis Function network as the decisive classifier, composed with the use of the constructivist Feature Space Mapping algorithm. In this configuration, the group correctly identified more than 99% of the elements of the validation set. The models may be used as tools for predicting expected soil condition, which is helpful in assessment of the range of substantial transformations.