Face To A Kind Of Retrieval Model Research Of Network Study

Currently, the individuation E-learning pattern is the most suitable for vast numbers of learners and the most popular E-learning pattern. One of its characteristics is exploration, that is to say, learners obtain needful learning resources by their own quest and study in the network. But with the development of information technology, the web learning resources present their infinity and complexity. The traditional keyword-based matched information retrieval technique can't meet the needs of learning activities of individuation and creative of learning habits from whether the covering rate or the retrieval accuracy of resources. Thus, this dissertation constructs a semantic-based retrieval model, puts forward a"cluster"-based RDF dynamic semantic retrieval algorithm, consequently solves three big problems: missing check, mismatch and low retrieval efficiency caused by traditional information retrieval technique.First, the dissertation analyzes and compares four E-learning patterns, and construct a semantic-based retrieval model for individuation E-learning pattern according to the characteristics of learning resources and their retrieval requirements of individualtion E-learning pattern. The model increases a semantic relation database on the basis of the traditional metadata-based resource retrieval model. The model makes use of the automatic constructed semantic relation database to realize semantic expansion (i.e. synonymy extension, intension extend, extension extend and parataxis extend) of E-learning resources metadata. Semantic relation database is the crucial part of the model. Therefore, a method is presented to automatically obtaining semantic relation database. Next, the dissertation introduces the applications of XML and RDF(S) in the network education, and uses XML/RDF(S) to carry through semantic description for E-learning resources metadata, so as to make computer can understand and process them. Then the dissertation mainly presents the RDF dynamic semantic retrieval algorithm, and on the basis of it, the dissertation puts forward a"cluster"-based RDF dynamic semantic retrieval algorithm. In the section, the dissertation describes in detail the model architecture of"cluster"-based semantic P2P network, the construction of super-peer semantic cluster and search arithmetic. Compared the"cluster"-based RDF dynamic semantic retrieval algorithm with original algorithm, the main virtue of the"cluster"-based RDF dynamic semantic retrieval algorithm is, when the query condition matches with the target resources, to restrict the search range to the attributive semantic clusters of super-peer and the presidial autonomic peer clusters of super-peer, that is to say, it reduces the search range from all peers in semantic P2P network to those peers of the attributive semantic clusters of super-peer and the presidial autonomic peer clusters of super-peer, avoids requested flooding effectively, greatly shortens search time and raises retrieval efficiency. Finally, the dissertation presents experiments as follows:1. The dissertation tests the influences of attenuation gene a on semantic relation pairs; Meanwhile, the dissertation tests the influences of semantic relation database constructed in the dissertation on information retrieval by adopting average query recall rate and average query accurate rate as evaluation standard.2. The dissertation tests the SemP performance by adopting query recall rate as valuation standard; Meanwhile, the dissertation tests the validity of"cluster"-based RDF dynamic semantic retrieval algorithm by adopting consumed times on query as valuation standard.