Efficient computation of a proximity matching in spatial databases

Spatial data mining recently emerges from a number of real applications, su ch as real-estate marketing, urban planning, weather forecasting, medical i mage analysis, road traffic accident analysis, etc. It demands for efficien t solutions for many new, expensive, and complicated problems. In this pape r, we investigate a proximity matching problem among clusters and features. The investigation involves proximity relationship measurement between clus ters and features. We measure proximity in an average fashion to address po ssible non-uniform data distribution in a cluster. An efficient algorithm i s proposed and evaluated to solve the problem. The algorithm applies a stan dard multistep paradigm in combining with novel lower and upper proximity b ounds. The algorithm is implemented in several different modes. Our experim ent results not only give a comparison among them but also illustrate the e fficiency of the algorithm. (C) 2000 Elsevier Science B.V. All rights reser ved.