Research On Semantic Retrieval For Communication Ontology

With the development of social information, the information growth explosive,how to get more accurate and effective useful information from huge amounts of data has been attracted more and more attention. Traditional retrieval is based on keywords matching characters, this method just output the result which matched the keywords.So, it will miss a lot of useful information, resulting in the recall and precision rate will not high. Semantic retrieval is based on the retrieval model of semantic information, this model has the analysis of the semantic meaning of the keywords,it can be accurately and efficiently improve the results which users wants to. Therefore,semantic search become a focus of information retrieval.After study of the principles, methods and tools for ontology construction, we using the seven-step method which is the highest maturity and using Protege4.3 build communication domain ontology. Base on the China Library Classification and keywords, we have got the relationship between concept set and concept in the field of communications,(synonymous relationships, epistatic relationships, the next relationship, the first family relationships, etc.) and stored to ontology with the My SQL database. The system resources mapping model has mapped from ontology to relational databases, which including from the body of the class, instance, the data type of the attribute value to the table in the database, records, data type mappings.It used Jena inference engine speculated semantic relationships between concepts,the inference rules both have Jena and self-constructed rules.Using an improved based on semantic similarity algorithm which added the edge type, quantify semantic distan-ce between concepts in the concept collection, and sorted according to quantify result-s of the concepts collection. SPARQL query the instance database to get the requirem-ents of instances.We used ICTCLAS conducted segmentation on the user’s input for POS tagging and remove the disable word of segmentation results, to facilitate the expansion concept.Realized the semantic retrieval system in ontology of communication domain.The system describe semantic information of the concept and semantic relationships between different concepts through the ontology, semantic retrieval system semantic extension and semantic reasoning skills. Experiments show that the semantic retrieval system based on ontology in recall and precision rates higher thantraditional keyword-based retrieval.