University of Houston

Finding interesting patterns in spatial datasets is essential for many applications. Spatial datasets have unique characteristics, for instance autocorrelation, the continuous nature of space, complex spatial data types, the importance of maps as summaries, and the necessity to deal with a large number of potential patterns. However, traditional data mining techniques do not take the unique characteristics of spatial datasets into consideration; consequently, they do not perform well for applications of this kind. Density-based data mining techniques showed advantages over other types in coping with the challenges of spatial data mining. The proposed research centers on the development of spatial data mining techniques that works on the top of density functions. Density estimation techniques for spatial data mining will be investigated, particularly supervised density estimation techniques that additionally consider a non-spatial variable of interest. Novel density-based spatial clustering algorithms will be designed and implemented. DCONTOUR a novel clustering algorithm that combines supervised density estimation techniques with contouring algorithms has already been developed in our preliminary work. DCONTOUR uses contour polygons to represent cluster models. The use of contour clustering for change analysis in spatial datasets is another theme of the proposed research. Finally, techniques that rely on contour polygons for cluster visualization and for deriving decision boundaries for classification algorithms will be investigated.