SEDL / STP / STP1017-EB / STP18625S

An Interactive Inquiry System for Materials Data Bases Using Natural Clustering

Pilgrim, RA
Senior staff scientists, SRS Technologies, Huntsville, AL

Johnson, PM
Senior staff scientists, SRS Technologies, Huntsville, AL

Falco, PM
Materials development engineer, Air Force Wright Aeronautical Laboratories, Wright-Patterson AFB, OH

Pages: 9    Published: Jan 1989

Download this paper for $25 PDF (144K)          View License Agreement
Abstract

User inquiry in a data-base management system (DBMS) takes on patterns that can be recognized by a monitoring system using hierarchical conceptual clustering. The application of natural clustering techniques to interactive query sessions permits the automatic development of predictive tools to assist the unsophisticated data base user. This paper defines the desired attributes of such a monitoring inquiry support system, discusses system design considerations, and provides an algorithm for the natural clustering of both numeric and nonnumeric property information. The implementation of such a system is described based on its application to the Spacecraft Materials Data Base SRS Technologies is currently developing.