Ontology-based Information Retrieval Service Research Of Text Resources

With the constant expansion of the internet, information resources on the network aregrowing exponentially. Currently, the number of web pages on the internet is reached to10billion，and millions pages more every day. Problem faced by people is not a lack ofinformation, but how to retrieve useful information from such amount of information.Information retrieval technology eases the demand of access to informantion to a certainextent.Traditional information retrieval techniques were based on match of words, phrasesbased on the syntax and were implemented using inverted index or directory. These retrievaltechniques are not only easy to use, quickly to get information, but also easy to implement.However, these traditional information retrievals also exist some problems. It is difficult forusers to express their intention by several words when retrieving information. Moreover,Because of keword matching, polysemy or a meaning with a variety of representations makesretrival results not to meet the users’ demands. Furthermore, Systems based on thesetechniques are difficult to express semantic information between concepts.This article first summarized problems the traditional information retrieval techniquesexisted, elaborated features and advantages of ontology-based text resource informationretrieval and also narrated domestic and national research progress for ontology-basedinformation retrival techniques.Secondly, informantion retrieval model of text resources based on ontology wasproposed. Key functional blocks of the model were detailed descripted. Some keytechnologies were proposed such as ontology concept semantic similarity calculation, accessto title and abstract of document, the creation of ontology.Thirdly, some open sources tools that would be used to implement the ontology-basedinformation retrieval system were introduced, such as Lucene, IKAnalyzer, Jena, Protégé.At last, By using open source tools, a information retrieval application of Java courcewas designed combined with the proposed model. A ontolgy of Java cource was created usingconcepts in Java cource as material. Implicit information and relationships between conceptscould be trenched by reasoning on the ontology. Queries were expanded combined withontology concept semantic similarity calculation, which has improved the retrieval efficiency.By comparing the proposed information retrieval model with traditional information retrievaltechnologies, we can see that the proposed model has a higher recall rate and precision rate.