SBIR-STTR Award

A clustering approach is proposed for identifying candidate reusable regions in any multi-agent decision support system. Clustering allows an axiomatized ontology to be analyzed in the context of its usage instead of in the context of its declaration. Pragati?s Multi-ViewPoint Clustering Analysis (MVP-CA) Technology facilitates analysis of semi-structured systems by clustering meaningful information entities. Manipulable reusable software artifacts such as model fragments and templates can be extracted out of existing systems through clustering, which can be subsequently used to create new ontologies and axioms. Both ontological and axiomatic aspects of legacy systems are thus addressed for reuse. The Façade Creation Toolkit (FCT) project provides an environment for addressing the creation and reuse of ontologies. Our joint approach promises to reduce the cost of developing multi-agents substantially as we propose an informed, interactive, adaptive and holistic approach for reusing legacy multi-agent systems. We anticipate that our proposed infrastructure can benefit several ontology authoring environments, especially the UML-based ones. In addition we anticipate that our research will have impact on the knowledge management market as well. Mining for knowledge in senmi-structured systems has applications in various fields such as forensics, bio informatics, criminal investigation, anti-terrorism-related analysis and so on.

Keywords:
Model Fragments, Clustering, Reuse Potential, Perspectives, Templates, Knowledge Mining, Ontology, Abstraction

In this Phase II project we propose to develop a clustering-based Ontology Search and Reuse Tool (OSRT) that enables builders of knowledge-based systems to identify and reuse relevant portions of existing systems, thereby speeding delivery, reducing development costs, enhancing the quality of the developed system, and, ultimately, increasing return on investment (ROI). Currently, aligning, mapping and merging knowledge-based systems are complicated undertakings that require significant manual effort from their designers. Three primary barriers to effective reuse are: where is the concept to reuse, how is it used currently, and how can it be adapted for reuse in the new application. OSRT will provide support on all these three fronts by acting as a semantic mediation layer across varios DOD and commercial stove-pipe applications that need to transition to new platforms.

Keywords:
Clustering, semantic mediation, reuse, interoperability, knowledge discovery, information extraction, knowledge-based systems