Mongolian Semantic Retrieval Research Based On Domain Ontology

With the vigorous development of the Internet, the information resources are growing rapidly, timely access to effective information has become an important force to promote social development and progress. Now the mainstream retrieval technology is in the index file to query word machinery matching query and does not consider the word itself has the semantic information query, so its retrieval result is bound to be existing in the natural language expression differences (such as synonyms) and lexical island effect, unable to meet the user’s demand. At present, there are already a part of the operating system and application software support Mongolian, along with the development of Mongolian information processing and Mongolian literature digital technology, Mongolian digital resources from scratch. However, the retrieval service of these resources is basically based on key words and string matching. Based on Mongolian domain ontology semantic retrieval breaks through the single mechanical matching word retrieval query method, the use of ontology specific semantic reasoning techniques to get the hidden information, rich query results to improve retrieval effectiveness.Firstly, this paper makes a research in ontology construction and semantic retrieval home and abroad, based on the proposed using ontology technology of semantic retrieval, semantic field of Mongolian music retrieval system framework is designed. The detailed research and describes the main techniques and tools for semantic retrieval based on Mongolian domain ontology. The main technologies include construction of domain ontology and semantic analysis and reasoning, the auxiliary tools include ontology modeling tool-Protege, semantic analysis and inference tool-Jena and so on. The procedures of Mongolian semantic retrieval are as follows:First, taking Mongolian music fields as an example and using Protege to construct domain ontology; second, using Jena to make an elaborate semantic parsing and inferencing, also making a detail introduction in the important components of the system and semantic retrieval process; finally, integrating key processes and then completing the testing system, also compared the results before and after the reasoning. This thesis will somewhat lay a foundation of application in Mongolian semantic web and enrich the semantic web field.