ANALYSIS OF PREFERENCE MAPS USING DATA MINING METHODS

Preference maps allow people to express their opinions about future spatial development in a simple way; they mark the areas that they consider suitable for specific activities on a cartographic base map. The decision on the area is generally intuitive and reflects their views and preferences regarding the solution of spatial problems. In this way, preference maps may hold valuable information and convey hidden knowledge. To make the best of their potential usefulness in spatial planning and to contribute to the transparency of the process, these information and knowledge should be explicit and presented in an interpretable form. In this paper, we report on a case study of preference maps analysis for municipality Komenda in Slovenia, where residents marked areas they considered especially valuable and which should therefore be preserved. In an attempt to discover why specific areas were marked for protection, we used the selected data mining approaches to infer the relations between preferential annotations and spatial characteristics. The inferred patterns were reported in the form of decision rules and in the graphical form of a nomogram. Interpretation of results shows that the methodology proposed in this paper and the explicit decision criteria and rules extracted by data mining can be useful for further applicability in spatial planning.