Dr. Samuel Singapogu received his PhD in Computer Science December 15, 2015. His dissertation describes an automated method for knowledge discovery and ontology matching using XML schemata.
“COMMAND AND CONTROL KNOWLEDGE DISCOVERY AND ONTOLOGY MATCHING USING XML SCHEMA”
Dissertation Director: Dr. Mark Pullen, Department of Computer Science
Committee: Dr. Paulo Costa, Dr. Robert Simon and Dr. Elizabeth White
Acknowledgement of gratitude to: Dr. Mark Pullen, Nick Clark, Priscilla McAndrews, Deb Schenaker, Jackie Kang and all the C4I staff for the support.
ABSTRACT
Samuel Singapogu, Ph.D.
George Mason University, 2015
Dissertation Director: Dr. Mark Pullen
Semantic modeling, which is necessary for semantic interoperability, formalizes domain knowledge. Semantic interoperability is crucial to Command and Control (C2) systems especially in a coalition environment where heterogeneous systems need to exchange digitized C2 data under shared semantics. Command and Control systems to Simulation systems Interoperability (C2SIM) is an evolving standard to represent and exchange digitized C2 and simulation data between heterogeneous systems such as C2 systems, simulation systems and robotic systems. C2SIM standard is based in part on XML based standardization of Coalition Battle Management Language (C-BML).
This dissertation describes an automated method for knowledge discovery and ontology matching using XML schemata. It includes a semantic mapping from XML schema elements to ontological elements and a novel method to use Part of Speech (POS) tagging of XML schema annotations to create and enrich semantic domain relationships. A contribution of this dissertation is a process and framework to perform ontology matching and alignment between two ontologies using domain specific XML metadata. This enables an analysis of correlation between competing or complementary ontologies. The output of ontology matching can be used as a semantic mapping between the two systems for coalition testing and translation. Another significant contribution of this dissertation is to present a metric that captures an index of semantic correlation between an ontology and domain text.
The knowledge discovery method is applied to the C-BML XML schema to create a C2SIM ontology. Subject Matter experts evaluated the C2SIM ontology; results of the evaluation are presented. Ontology matching using XML schemata is applied to subset ontologies created from the C2SIM ontology. The results of ontology matching in comparison to the edit-distance method are provided. The use of index of semantic correlation is presented by comparing a domain document snippet to the C2SIM ontology.