Classification Of Dispersal Areas Using A Geographic Factor Application To Dialectology

We propose a multidimensional statistical analysis procedure using projection and classification methods, in order to identify coherent clusters into a set of surficial entities called dispersion areas. The methodology includes a geographical factor to build the representation space for the projection of the data. By applying this method on geolinguistic data, we are able to identify and explain new spatial patterns among a set of dispersion areas of linguistic features.