Advanced search
1 file | 1.75 MB

Uncertainty propagation in vegetation distribution models based on ensemble classifiers

Jan Peters (UGent) , Niko Verhoest (UGent) , Roeland Samson, Marc Van Meirvenne (UGent) , Liesbet Cockx (UGent) and Bernard De Baets (UGent)
(2009) ECOLOGICAL MODELLING. 220(6). p.791-804
Author
Organization
Abstract
Ensemble learning techniques are increasingly applied for species and vegetation distribution modelling, often resulting in more accurate predictions. At the same time, uncertainty assessment of distribution models is gaining attention. In this study, Random Forests, an ensemble learning technique, is selected for vegetation distribution modelling based on environmental variables. The impact of two important sources of uncertainty, that is the uncertainty on spatial interpolation of environmental variables and the uncertainty on species clustering into vegetation types, is quantified based on sequential Gaussian simulation and pseudo-randomization tests, respectively. An empirical assessment of the uncertainty propagation to the distribution modelling results indicated a gradual decrease in performance with increasing input uncertainty. The test set error ranged from 30.83% to 52.63% and from 30.83% to 83.62%, when the uncertainty ranges on spatial interpolation and on vegetation clustering, respectively, were fully covered. Shannon’s entropy, which is proposed as a measure for uncertainty of ensemble predictions, revealed a similar increasing trend in prediction uncertainty. The implications of these results in an empirical distribution modelling framework are further discussed with respect to monitoring setup, spatial interpolation and species clustering.
Keywords
Spatial interpolation, Distribution model, Vegetation type, Random Forests, Ensemble learning, Classification, Uncertainty, Wetland, Clustering

Downloads

  • (...).pdf
    • full text
    • |
    • UGent only
    • |
    • PDF
    • |
    • 1.75 MB

Citation

Please use this url to cite or link to this publication:

Chicago
Peters, Jan, Niko Verhoest, Roeland Samson, Marc Van Meirvenne, Liesbet Cockx, and Bernard De Baets. 2009. “Uncertainty Propagation in Vegetation Distribution Models Based on Ensemble Classifiers.” Ecological Modelling 220 (6): 791–804.
APA
Peters, Jan, Verhoest, N., Samson, R., Van Meirvenne, M., Cockx, L., & De Baets, B. (2009). Uncertainty propagation in vegetation distribution models based on ensemble classifiers. ECOLOGICAL MODELLING, 220(6), 791–804.
Vancouver
1.
Peters J, Verhoest N, Samson R, Van Meirvenne M, Cockx L, De Baets B. Uncertainty propagation in vegetation distribution models based on ensemble classifiers. ECOLOGICAL MODELLING. 2009;220(6):791–804.
MLA
Peters, Jan, Niko Verhoest, Roeland Samson, et al. “Uncertainty Propagation in Vegetation Distribution Models Based on Ensemble Classifiers.” ECOLOGICAL MODELLING 220.6 (2009): 791–804. Print.
@article{517670,
  abstract     = {Ensemble learning techniques are increasingly applied for species and vegetation distribution modelling, often resulting in more accurate predictions. At the same time, uncertainty assessment of distribution models is gaining attention. In this study, Random Forests, an ensemble learning technique, is selected for vegetation distribution modelling based on environmental variables. The impact of two important sources of uncertainty, that is the uncertainty on spatial interpolation of environmental variables and the uncertainty on species clustering into vegetation types, is quantified based on sequential Gaussian simulation and pseudo-randomization tests, respectively. An empirical assessment of the uncertainty
propagation to the distribution modelling results indicated a gradual decrease in performance with increasing input uncertainty. The test set error ranged from 30.83\% to 52.63\% and from 30.83\% to 83.62\%, when the uncertainty ranges on spatial interpolation and on vegetation clustering, respectively, were fully covered. Shannon{\textquoteright}s entropy, which is proposed as a measure for uncertainty of ensemble predictions, revealed a similar increasing trend in prediction uncertainty. The implications of these results in an empirical distribution modelling framework are further discussed with respect to monitoring setup, spatial interpolation and species clustering.},
  author       = {Peters, Jan and Verhoest, Niko and Samson, Roeland and Van Meirvenne, Marc and Cockx, Liesbet and De Baets, Bernard},
  issn         = {0304-3800},
  journal      = {ECOLOGICAL MODELLING},
  keyword      = {Spatial interpolation,Distribution model,Vegetation type,Random Forests,Ensemble learning,Classification,Uncertainty,Wetland,Clustering},
  language     = {eng},
  number       = {6},
  pages        = {791--804},
  title        = {Uncertainty propagation in vegetation distribution models based on ensemble classifiers},
  url          = {http://dx.doi.org/10.1016/j.ecolmodel.2008.12.022},
  volume       = {220},
  year         = {2009},
}

Altmetric
View in Altmetric
Web of Science
Times cited: