Research And Application Of Semantic Retrieval Based On Course Ontology

With the rapid development of computer network technology, E-learning, which is a teaching model not restricted by time and space, causes great attention of educators. And E-learning has become an important supplement of traditional teaching. As an important part of Web-based teaching, the retrieval of teaching information has become main way by which learners get knowledge resource. Most of information retrieval is based on key word literally matching, such as Boolean Model,Vector Space Model and so on. These retrieval models can meet learners'need to a certain exten, but they have limitation on information semantics. And they are also lack of processing and understanding of knowledge.Semantic search is a kind of knowledge-based,semantic analysis retrieval. It is based on semantic understanding, and gives semantics to users'search request,information resources of retrieval and results of retrieval. Semantic search uses concept matching mechanism to search. Because of the lack of keyword information retrieval, combining the project of Digital Education Publishing Support Platform of Hi Education Press, researching on the technology of ontology and information retrieval, this paper proposes a semantic retrieval model based on course domain ontology. This paper applies domain ontology to analysis and process the semanteme of query sentences and document resources, then understands the query need and the latent semantic information of document resources, in order to establish semantic vector and achieve semantic retrieval.The main work of this paper includes: research and create course ontology,Chinese word segmentation,research on semantic retrieval model,design and implement retrieval system. And this paper focuses on ontology and semantic retrieval model.1) Combining the characteristics of the current teaching resource library, this paper designs knowledge hierarchy architecture of teaching resource library. This architecture provides theoretical basis for level of course domain ontology concept in order to achieve the mapping of course knowledge to domain ontology. This paper extracts concept and relation of《Principles of Computer Composition》course knowledge, and then researches and applies the inference mechanism of ontology, in order to provide semantic basis for retireval based on ontology. 2) This paper uses Chinese lexical analysis system ICTCLAS to achieve word segmentation of document resources and query sentences. Additionally, this paper looks course domian ontology as knowledge base dictionary to achieve the recogintion of professional word.3)The semantic retrieval algorithm that this paper proposes bases on Vector Space Model. Accroding to the concept and relation of course domain ontology and the results of inference to calculate the semantic weights, in order to achieve semantic reconstruction of document resources and semantic extension of query sentences. And the goal is to achieve the semantic information retrieval. Combining with the reality of teaching, retrieval model extends query word by different semantic relation in the process of vectorization of query sentences. So the results of retrieval are more in line with teaching needs.4)According to the project of Hi Education Press, this paper designs and implements the semantic retrieval system. This system can get latent semantic information of document resources and understand the need of query request, and then returns teaching information resources which meet query need. In the end, this paper compares and analyses the advantage of proposed semantic retrieval algorithm and traditional retrieval algorithm. The experiment shows that new algorithm can do well in raising the rate of recall and precision.