/spl alpha/-surface and its application to mining protein data

Given a finite set of points in three dimensional Euclidean space R/sup 3/, the subset that forms its surface could be different when observed at different levels of detail. In this paper we introduce a notion called /spl alpha/-surface. We present an algorithm that extracts the /spl alpha/-surface from a finite set of points in R/sup 3/. We apply the algorithm to extracting the /spl alpha/-surfaces of proteins and discover patterns from these surface structures using a pattern discovery algorithm. We then use these patterns to classify the proteins. Experimental results demonstrate the good performance of the proposed approach.