Knowledge Discovery from Geographical Data

Abstract

During the last decade, data miners became aware of geographical data. Today, knowledge discovery from geographic data is still an open research field but promises to be a solid starting point for developing solutions for mining spatiotemporal patterns in a knowledge-rich territory. As many concepts of geographic feature extraction and data mining are not commonly known within the data mining community, but need to be understood before advancing to spatiotemporal data mining, this chapter provides an introduction to basic concepts of knowledge discovery from geographical data. In performing knowledge discovery in a spatial data set, the first important question is how to use the spatial dimension in the discovery process. At least two viewpoints can be considered: either spatial relationships are made explicit prior to data mining or specialised algorithms are directly applied to spatial and nonspatial data. The first approach claims that the spatial dimension is somewhat more basic than the other features, and, then, it can be used to prepare the data set for a successive knowledge extraction step. The exploitation of the spatial dimension for selecting the values of attributes to be used in the mining step can be quite complex, and it may depend both on the structure of the domain and on the kind of model one is looking for. This approach offers the advantage of allowing the reuse of standard data mining technology on data extracted according to the spatial dimension. The second approach aims at exploiting the spatial features dynamically during the discovery process. This implies a complete reinvention of the data mining technology, but it allows a more flexible use of spatial knowledge. Mining geographic data poses additional challenges, which include the exploitation of background knowledge as well as the handling of spatial autocorrelation