Multi-ViewPoint Clustering Analysis for Validation of Knowledge-Based Systems
  • Agency: NSF (Phase I SBIR)
  • Duration: January 1997 - June 1997.

   Capabilities Being Demonstrated:
  • Exposing the software architecture
  • Exposing Verification & Validation
  • Exposing Reusable Regions in Software

Pragati was awarded a Phase I SBIR from National Science Foundation to implement a portable and user-friendly prototype version of the MVP-CA tool. Pragati developed a Java/Web based version of the tool that could generate multiple clusterings on an expert system within a single execution. A new user interface was designed and implemented using the Java Abstract Window Toolkit (AWT). The use of Java made the tool portable across architectures including Sun Solaris, PC based Windows 95/98 and the Macintosh.

The tool can be invoked as a stand-alone application or through a browser interface allowing the tool to be used remotely across networks. The tool was extended to utilize the multi-threading capabilities of Java. This allows users to analyze a system from different viewpoints in different windows simultaneously while still sharing the underlying data structures. This capability can easily expose different legitimate perspectives of a knowledge-based system and make quick comparisons across viewpoints possible.

  1. M. Mehrotra and V. Barr. Application of Clustering Analysis to Coverage Testing for Large KBS. Workshop on Validation, Verification and Testing of Knowledge-Based Systems in the AAAI Conference, July 1998, Madison, Wisconsin.

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