Polychotomous regression : application to landcover prediction

Frédéric Ferraty (1) , Martin Paegelow (2) , Pascal Sarda (1)

1 IMT - Institut de Mathématiques de Toulouse UMR5219
2 GEODE - Géographie de l'environnement

An important field of investigation in Geography is the modelization of the evolution of land cover in view of analyzing the dynamics of this evolution and then to build predictive maps. This is possible with the apparatus of measure : sattelite image... In this paper, we propose to use a polychotomous regression model to modelize and to predict land cover of a given area : we shox how to adapt this model in order to take into account the spatial correlation and the temporal evolution of the vegetation indexes. This study concerns an area in the Pyrenees mountains.

Mots clés: Polychotomous regression; land cover prediction; Pyrenees; mountains

Domaines: Géographie

Frédéric Ferraty, Martin Paegelow, Pascal Sarda. Polychotomous regression : application to landcover prediction. Statistical case studies, Springer Verlag, 13 p. en ligne, 2005. ⟨halshs-01062315⟩