Grupowanie metodą k-średnich z warunkiem spójności
Clustering using k-means method with coherence property

Summary/Abstract: Clustering methods are often used to group administrative units (municipalities, counties, regions, countries). Their most common subjects are: quality of life, investment attractiveness of regions, the level of development of infrastructure or other directly immeasurable phenomena. In the classic procedures for clustering one does not consider the spatial relationships of administrative units, which can lead to some cognitive dissonance when the isolated groups are characterized by the lack of spatial coherence. In the paper the modified algorithm of the k-means method with coherence property and its implementation in the form of an extension of the STATISTICA software was presented. Some examples of clustering results into coherent subsets for the data on districts were presented. The obtained division was compared to the classic division of the classification, which does not take into account the coherence condition.

• Details
• Contents