Application of Multi-ViewPoint Clustering Analysis to an Expert Systems Advocate Advisor
  • Agency: Department of Transportation.
  • Duration: December 1994 - March 1995.

   Capabilities Being Demonstrated:
  • Extraction of the Software Architecture
  • Reduction of Test Cases
  • Exposing Restructurable Regions in Software
  • Verification & Validation
    • Conflict or rule pair anomalies
    • Redundancy
    • Incompleteness

In 1994 Pragati was awarded a follow-on contract for applying the MVP-CA technology on a small, but poorly structured knowledge base, the Expert System Advocate's Advisor (ESAA) that was also known to have multiple errors. The MVP-CA tool surprised its developers by discovering the underlying software architecture of the knowledge base. Since the knowledge base had been developed in a very ad hoc manner, none of the developers had any idea of the interconnections among the concepts. Thus MVP-CA proved its usefulness during the evolving state of a knowledge base when multiple experts update the knowledge base with their own expertise. Having a (semi-) automated means of exposing the current underlying structure can pave the way towards making intelligent architectural design decisions during development. Also, Pragati was able to discover several instances of inconsistent and redundant rules in ESAA, which were automatically detected by the MVP-CA tool. Incomplete specifications on various variable and qualifier values surfaced easily as well through the clustering of rules from multiple perspectives by the MVP-CA tool.

  1. M. Mehrotra. Application of Multi-ViewPoint Clustering Analysis to an Expert Systems Advocate Advisor. Technical Report, Publication No. FHWA-RD-97-022. Federal Highway Administration, McLean, VA, November 1997.

« Back | Top
2007 Pragati Synergetic Research, Inc. Research Focus | Products | Projects | Publications | Market Sectors | About Us