Ontology Semantic Information Retrieval Based On Natural Language Understanding

Traditional information search techniques are based on keyword matching and full-t-ext query syntax technology, but there are many problems, for example:"faithfully expr-ess", mechanical matching problem, "vocabulary island" problem, can't accurately reveal the essential contents of information etc. Semantic search is a knowledge-based, semantic search, semantic understanding, based on the user's search request, search and retrieve th-e results of the object resources are given a certain semantic, can effectively improve the precision.Based on the existing basis of semantic retrieval, combined with a variety of diction-aries and natural language processing technology, the ontology semantic information retr-ieval based on natural language understanding was put forward. Firstly, the synonyms ex-traction rules was defined according to the definition of schema, gain the domain ontolo-gy synonym from the relevant information (eg Baidu Encyclopedia). Followed by Jena inference engine, defing reasoning rules according users themselves and combine with domain knowledge to obtain implicit knowledge. Word similarity algorithm was propose-d based on Synonym word Lin, natural language query entered by the user was syntax a-nalyzed, semantic analyzed, and SPARQL query was constructd to achieve the natural L-anguage queries. Bias on the accuracy of formula was proposed in order to measure the r-easonableness, and the content of user interaction was analyzed. Finally, in the query res-ults returned a wealth of information to help users understand the contents of the query a-nd the achievement of the second query.This example implements the semantic ontology of computer information retrieval, experiments show that the method can access semantic information, to effectively impro-ve the system precision.