Meta-modeling for Multi-Model Integration
Date
2010-06-08T13:41:46Z
Authors
Rafi, Muhammad Faraz
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
This thesis work proposes a technique based on concept maps, meta-models, and ontologies to develop a theoretical foundation for the use of multiple interacting models in order to determine the valid interaction between different modeling techniques. It can assist a modeler to find out if a model can provide partial answers to queries generated by another one. The modeling techniques considered in this thesis are Social network and Influence net. A phase by phase approach using concept maps, meta-models, and ontologies based on comparing the ontologies (for each modeling technique) is employed which helps identify the similarities, overlaps, and/or mapping across the models. In the first phase, concept maps are developed for each technique by identifying a set of focus questions. This concept map representation is then formalized using meta-models. The aim of constructing the meta-model is to reveal the structural aspects of the modeling technique and lay down the foundation for its ontology. A basic ontology is constructed which mirrors the meta-model and serves as the foundation ontology; it does not contain any explicit concepts (related to the modeling technique). As the next step, explicit concepts and relationships are added to this foundation ontology to make it complete. Once the individual ontologies are completed, mapping of concepts across the ontologies is performed. The resulting ontology which contains these explicit concepts and relationships within and across both ontologies is called as enriched ontology. The mappings determined with the help of this enriched ontology can then pave the way towards ensuring consistency and exchanging information among the two different types of models. An example based on an actual event is used for the construction of a Social network and an Influence net model. The construction of both Social network and the Influence net is based on the same corpus of data, but they are constructed in isolation. A domain specific ontology for this case study is instantiated from the enriched ontology. The mappings determined with the help of the enriched ontology help translate the analysis output of the Influence net model and incorporate it into the corresponding Social network. As a result of this incorporation changes made in the Social network can then be analyzed and studied for any revealing information about the network.
Description
Keywords
Meta-model, Ontologies, Concept map, Influence net, Multi-modeling, Social network