Knowledge discovery in spatial databases through qualitative spatial reasoning

Abstract

Human beings use qualitative identifiers extensively to simplify reality and to perform spatial reasoning more efficiently. Organisational databases usually store geographic identifiers, like addresses or postcodes, which spatial component is not incorporated in the knowledge discovery process. This paper addresses the process of Knowledge Discovery in Spatial Databases through a Qualitative Spatial Reasoning approach. The aim is the improvement of the referred process by the adoption of qualitative identifiers like North, South, close, far, etc. in the classification of spatial relations that exists between the geographic entities addressed. The proposed approach uses a spatial reasoning strategy that integrated direction and distance spatial relations in the reasoning process, allowing the inference of implicit spatial relations for the several levels of the considered geographic hierarchies. The integration of a geographic and a demographic database allowed the discovery of spatial patterns and general relationships that exist between the analysed spatial and non-spatial data.