Research On Retrieval Model Based On Ontology

With the rapid development of the network, the network has become the explosive growth of information. For the flood of information, more and more people are willing to through the use of search engines to find their own from the network information needed. However, traditional keyword-based search system that matches the user query language due to lack of understanding, often returned to the user a lot of irrelevant information. Therefore, the semantic retrieval is being more and more attention. Meanwhile, Proposal and development of the Semantic Web and Ontology also provide the foundation for the semantic retrieval research.In the study of traditional search engines, this dissertation presents a retrieval method based on ontology. In the research and understanding of the retrieval system based on ontology, this dissertation improves some traditional algorithm. And builds a semantic information retrieval system based on a simple domain ontology, experiments comparison with the traditional retrieval system.This research work mainly includes:1. According to the existing corpus, build a domain ontology of computer science. Domain ontology is created when the corpus is also considered in the ontology concepts and instances of co-existing factors. The ontology is constructed not only from the perspective of the field of domain science, but also from the statistical point of view.2. Pretreatment in the documents, the use of the traditional TF-IDF algorithm also considers the characteristics of words in the document location factors.3. When a user's query is semantics expanded, this dissertation is based on the ontology of semantic expansion, but also the local context analysis method used as a supplement to reduce the retrieval system's dependence on the domain ontology to improve the system's applicability.The experiment results obtained from the retrieval system to ensure the precision is not much lose, the recall rate has been greatly improved than the traditional retrieval system.As this dissertation is based on domain ontology for the semantic retrieval, the semantic search based on multi-domain and even generic ontology is not too much involved. The major direction will be my future research and work.