Ontology-based Semantic Retrieval Research

Traditional information retrieval is mainly based on keywords, they compares the user's retrieval request with each word in the context. Because the original meaning and the meaning extended of the concept is not in the same level, which it makes the search engine misunderstand the intention of the user accurately. As a result the solution only matches the literal meaning of the concept, what's more, the recall and precision ratios of the retrieval are not very quite high. With a good concept hierarchical structure and logical reasoning function support, the ontology can express semantic information based on relations between concepts. Therefore, ontology-based semantic retrieval can understand the user's query intention accurately and exploit semantic information which not only is relevant to the concept but also meet the user's need. The ontology can achieve retrieval in semantic level, which overcomes the limitations of traditional information retrieval technology.This thesis first introduces the state of study on ontology-based semantic retrieval at home and abroad; then it proposes the concept of semantic information chain, and describes the achievement in semantic information in ontology semantic retrieval, to achieve the semantic-level retrieval. The main research contents are as follows:Firstly, proposing the concept of semantic information chain, which based on the advantages of the ontology-based semantic retrieval:reflect semantic information. The semantic information chain includes three steps:ontology construction, semantic indexing and semantic retrieval. The semantic retrieval is the key.Secondly, introducing the basic process that reflects the semantic information of the semantic retrieval, including query expansion and the results sorted. Semantic retrieval of semantic information chain uses semantic information in ontology library first, and then expanses the semantic query into related concepts. Based on the user input mode, it also introduces the basic idea of query expansion, and proposes corresponding algorithm.Next, proposing a semantic retrieval algorithm. Based on receiving query concept, it matches the concept of query expansion with the semantic information of request contents, by calculating the directly similarity of concepts and documents, top the results consistent with the user's needs for the user searching, and through which it proposes a semantic retrieval algorithm.Finally, verifying and analyzing the algorithm. Based on C++Programming bilingual resources, a semantic retrieval system is developed to verify the effectiveness of the proposed algorithm.In this thesis, we make systematic research on the ontology-based semantic retrieval, propose a semantic expansion and a semantic retrieval algorithm, and also verify the effectiveness of them. How to more deeply achieve the semantic information in the process of semantic retrieval, and embody the retrieval algorithm more deeply, are our goals in the future.