Does vegetation suit our models ? Data and model assumptions and the assessment of species distribution in space

Auteurs :
BIO, A.M.F.
The Royal Dutch Geographical Society, Pays-Bas

Description :
The general aim of the research was to investigate ways in which empirical regression models that relate vegetation species distribution to site conditions could be improved. Some of the most elementary current problems in ecological logistic regression modelling are adresses (het Likelyhood Ratio Test, the Akaike's Information Criterion and the Bayesian Information Criterion). The results demonstrate the presence of serious problems in crucial parts of common modelling procedures, since model selection, parameter estimation and model evaluation are biased because of invalid assumptions. - (AGD)


Type de document :
Monographie

Source :
Collection Nederlandse geografische studies, issn : 0169-4839, 2000, vol. 265, nombre de pages : 195

Date :
2000

Identifiants :
isbn : 90-6809-288-X

Editeur :
Pays édition : Pays-Bas, Groningen, Faculteit der Ruimtelijke Wetenschappen Rijksuniversiteit Groningen