The usefulness of unsupervised classification methods for landscape typification: The case of Slovenia

Abstract

Supervised and unsupervised classification methods can be a useful tool in determining various geographical spatial divisions, especially regionalizations and typifications. Because Slovenia is geographically very diverse, its divisions are a particularly significant and interesting research challenge. The main objective of this article is to determine the effectiveness of unsupervised classification methods, and therefore we compare the well-established landscape typology of Slovenia from 1996 with landscape typologies that were modeled using various unsupervised classification methods. Our results show that landscape typologies modeled using unsupervised classification methods deviate more from the original landscape typology of Slovenia than landscape typologies modeled using random and expert-supervised classification methods.