Intelligent materials processing by hyperspace data mining

Abstract

This paper discusses application of hyperspace data mining to materials manufacturing. We introduce an innovative hyperspace method whereby data are separated into subspaces, features selected according to data patterns, and control rendered in the original feature space. This technique has three major advantages: no equipment is added, no experiment is needed, and no interruption occurs to production. A number of proprietary algorithms have been built into a software product, MasterMiner/sup TM/, for use in materials design and manufacturing. Examples are given to show the efficacy of the proposed method and MasterMiner tool.