Simplifying syntactic and semantic parsing of NL-based queries in advanced application domains

• Published in: Data & knowledge engineering Vol. 55; no. 1; pp. 38 - 58
• Main Authors: Kapetanios, E., Baer, D., Groenewoud, P.
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.10.2005
• Subjects: Automata theory; Information retrieval; Natural languages; Ontologies; Query languages; Semantic web; Semantic web; Databases; Natural languages; Automata theory; Ontologies; Information retrieval; Query languages
• ISSN: 0169-023X; 1872-6933
• DOI: 10.1016/j.datak.2004.11.008

The paper presents a high level query language (MDDQL) for databases, which relies on an ontology driven automaton. This is simulated by the human–computer interaction mode for the query construction process, which is driven by an inference engine operating upon a frames based ontology description. Therefore, given that the query construction process implicitly leads to the contemporary construction of high level query trees prior to submission of the query for transformation and execution to a semantic middle-ware, syntactic and semantic parsing of a query with conventional techniques, i.e., after completion of its formulation, becomes obsolete. To this extent, only, as meaningful as possible, queries can be constructed at a low typing, learning, syntactic and semantic parsing effort and regardless the preferred natural (sub)language. From a linguistics point o view, it turns out that the query construction mechanism can easily be adapted and work with families of natural languages, which underlie another type order such as Subject–Object–Verb as opposed to the typical Subject–Verb–Object type order, which underlie most European languages. The query construction mechanism has been proved as practical in advanced application domains, such as those provided by medical applications, with an advanced and hardly understood terminology for naive users and the public.