Font Size: a A A

Research On Multi-Source Linked-Data Ranking Using Rough Set

Posted on:2015-01-04Degree:MasterType:Thesis
Country:ChinaCandidate:W ZhangFull Text:PDF
GTID:2268330428967147Subject:Information Science
Abstract/Summary:PDF Full Text Request
Semantic Web provides us with a possibility to process web data with an intelligent approach, and linked data has become the main data type in the current semantic applications revealing a semantic web model with linked data datasets as central nodes. But as the number of datasets grow, problems of how to rank the resource acquired via querying multiple linked datasets have emerged. This ranking problem must be resolved before linked data can truly begin to elaborate its advantages regarding the aspects like openness and intelligence.Firstly, the article introduces the vital role that linked data plays in semantic web researches via those aspects:definition, application and ways to publish or query linked data. Thus raising the key problem in ranking the resource acquired via querying multiple linked datasets, which is, the difficulty to analyze and rank linked data resource from different datasets under a unified standard due to the variety of RDF vocabularies used in these datasets. Then the theory of rough sets is elaborated to establish an approach to complete the attribute reduction of integrated vocabulary using an algorithm based on attribute significance. By achieving the attribute reduction of integrated vocabulary which is the assemblage of all the RDF vocabularies related to each queried datasets, the algorithm provides a unified RDF vocabulary needed in multi-source linked data ranking to us. With this unified RDF vocabulary, a overall model of multi-source linked data ranking can be constructed. The article illustrated the overall model from three aspects: principle, layer model and ranking algorithm. Finally an experiment of this overall model and its ranking algorithm is conducted and the results shows that the model has achieved its designed purpose and possessing a certain level of practical potential.In this article Multi-source linked data ranking uses rough set theories to construct an overall ranking model based on a unified RDF vocabulary. This method has shown the capability to rank linked data in a complicated data environment on a sound basis. The research provides a possibility to transfer linked data query from single source to a multi-source environment and also shows an practical scheme for linked data applications to fit in the multi-source linked data query.
Keywords/Search Tags:Linked Data, Linked Data Ranking, Rough Set, Attribute Reduction
PDF Full Text Request
Related items