Semantic Querying with Ontologies | Saba Kashan Fallah
| | Abstract | In this project the usage of syntactic-semantic lexical resources like FrameNet combined with formal ontologies for the purpose of semantic interpretation of domain specific text, is researched. Accurate ontology driven semantic interpretation of natural language is only possible in terms of mappings based on reliable linguistic semantics, that maps linguistic forms to ontological types. The novelty of the method proposed in this report is to extend frame semantics with formal domain specific ontological types, in order to achieve deep semantic interpretation of text. The aim is deep semantic interpretation of scientific biomedical text, which in contrast to shallow semantic interpretation is not based on linguistic semantics only. Semantics of text is represented in terms of ontological concepts/classes, and frame-based definitions/descriptors. The twofold semantic representation provides not only mappings from text to nodes in ontological hierarchies and reverse, but also a mapping from query descriptors to ontological concepts and vice versa. Semantic annotations provided in the form of descriptors are used in order to semantically search text, where descriptors of user queries are matched against text-descriptors. Real ontological information retrieval is implemented in terms of descriptor subsumption based on the ontological hierarchies. Search based on the ontological hierarchy or input queries or a combination of both is made possible. The ontological structures are incorporated in the system, not only as a collection of terms but true to their nature as semantic networks. Methodologies and techniques for implementation of the logical core of such a multi-dimensional system, is suggested along with a proof-of-concept demo. | | Keywords | semantic interpretation, formal ontologies, frame semantics, ontology driven information retrieval, frame-based grammar, ontological grammar, parsing techniques | | Type | Master's thesis [Academic thesis] | | Year | 2007 | | Publisher | Informatics and Mathematical Modelling, Technical University of Denmark, DTU | | Address | Richard Petersens Plads, Building 321, DK-2800 Kgs. Lyngby | | Series | IMM-Thesis-2007-89 | | Note | | | Electronic version(s) | [pdf] | | BibTeX data | [bibtex] | | IMM Group(s) | Computer Science & Engineering |
|