Semantic Retrieval Of Semi-Structured Text Based On Ontology Concept

Accompany by the development of computer networks and knowledge society, a large number of semi-structured data showing on the Internet. Because of the characteristics of semi-structured data, the retrieval method of semi-structured data is different from the retrieval of structured data and unstructured data.Most of the semi-structured text retrieval system still use the traditional retrieval model based on keyword matching, which only match queries and documents in the literal, ignores the existence of the structural features and process the semi-structured text as a flat word sequence, therefore, the traditional retrieval based on semi-structured text has various deficiencies and it is impossible to improve the quality of search results, this dilemma need to be resolved.Semantic retrieval follows requirements that the retrieval system should be more intelligent and knowledgeable, it received a great deal of attention in academia. The semantic retrieval will solve the semantic gap that exists between people and computer. Concept retrieval is a new direction of the semantic retrieval, it makes the retrieval system broke through the literal match limitation and up to understand and deal with the user’s request from the significance concept level.This paper proposed a set of solution for semi-structured text semantic retrieval based on ontology concept, the solution take the concept correlation calculation based on ontology, the setting of term weight based on text structure, semantic query expansion, and semantic concept retrieval model into account. The experiment results show that the solution is feasible, effective and valuable in theory and application.