Density Data Generation for Spatial Data Mining Applications
23 August 2000
This paper describes a way to produce a particular type of density information from spatial data for spatial analysis applications or spatial data mining. The specification of an abstraction resolution for a chosen spatial framework is all that is required to produce this density information. This detailed spatial framework used for the production of the density information should also be that used for the subsequent spatial analysis. What is good about this way of producing density information is that once a spatial framework for the spatial analysis has been detailed, no further subjective bias is imposed on the resulting density information (even from NODATA space). In other words, the spatial bias is implicit to the shapes of the frame used to tessellate data space. The focus here is the production of density information for two-dimensional square grids or rasters containing NODATA space that are commonly use in geocomputation.
Reference: Proceedings of the Fifth International Conference on GeoComputation, University of Greenwich's School of Earth and Environmental Sciences, Kent, UK, 23 - 25 August 2000. Papers published on CD-ROM. Produced by: R.J.Abrahart and B.H.Carlisle. Publisher: "GeoComputation CD-ROM". ISBN 0-9533477-2-9