Ontology-Based Domain Resource Semantic Retrieval

Information retrieval technology has been profoundly influencing our working,study and life. With the popularization of internet, new higher demands have beenraised. Semantic retrieval based on conceptual matching is the combination of IRtechnology and artificial intelligence, and which can concern more about the querydomain and pay more attention to the semantic relationships between query words andretrieval targets. The introduction of ontology is for its hierarchical structure andrelations among concepts, and for its superiority in reusing, sharing, and knowledgeorganizing. This paper begins with the discussing of ontology concept, currentsituation of ontology research at home and abroad, and ontology in semantic web. Forconvenience, the study was divided into three stages: domain ontology construction,metadata extraction, and the establishment of retrieval model, which is the guiding ofthe study. Learning from early experiences and with actual needs a simplemethodology was developed for domain ontology construction, based on which aontology was built as an example and was coded in OWL. For conceptual matchingand semantic enhancement, the idea of semantic indexing was founded based on theacquisition of concept/instance and values of properties. For semantic similaritycalculating, a feasible formula, the combination of concept similarity and propertysimilarity, was defined. Then, the architecture of ontology based domain resourceretrieval system was proposed, with five central functional modules, and a retrievalalgorithm was designed to describe the retrieval process and to make a proof of theretrieval system in the formalizing way. At last, on basis of the semantic retrievalmodel, an experimental study was presented as verification and evaluation in someindicators of statistics.