Research On Strategies Of Vocabulary Reuse For Linked Data

With the increasing use of linked open data,data providers not only published their datasets as LOD,but also need to model the datasets in an easy-to-handle manner during the publishing process,making the data more user-friendly.So that it's easy for users to understand,query and use.And easy for computers and linked data applications to process.One of the best practices for linked data is to reuse existing vocabularies during the data modeling process,that is to reuse classes and properties from existing vocabularies,and use authority terms to describe data or to link the different datasets through these terms.Reusing vocabularies can improve the interoperability of datasets and avoid unnecessary waste of resources.At the same time,the semantic consistency of linked data and the existing datasets can be ensured to the maximum extent,so that it can be directly consumed by the general linked data applications without mapping and other processing.In practice,however,due to the increasing number of reusable vocabularies and the uneven quality of vocabularies,data publishers are unable to have a complete grasp of all vocabularies and the terms within the vocabularies.Besides,because of the professionalism of the data publishers themselves,it's difficult to decide exactly and quickly which vocabularies to describe the semantics of data.Therefore,the service demand of how to reuse vocabularies is increasing.In the data modeling process,how to determine the strategies of reusing vocabularies according to the needs of data publishers that involves many factors,such as the number of vocabularies,its compatibility with the modeling fields,its popularity or generality,the organization that published it and its maintenances.Therefore,this paper studys the strategies of linked data vocabularies reuse with theoretical discussion and empirical research,mainly focuses on the following aspects:(1)Introducing four principles and three characteristics of linked data,and it discusses the publishing process of linked data in detail.Analyzing the necessity of reuse vocabularies in the process of data publishing from the aspects of data preparation and definition of URI,choice of vocabularies,construct of RDF links,Web publishing and open query of data.(2)Defining the definition of vocabulary,and it summarizes and introduces the query service of vocabularies,then puts forward the vocabulary reuse process of linked data.And discusses the various influencing factors of vocabularies reuse,including the quality of vocabularies,management of vocabularies,data publishers and so on.(3)Through the introduction of linked open data cloud datasets to introduce the publication datasets.The datasets are classified and statistically analyzed for its vocabulary reuse category,influencing factors and RDF predicates.Then it analyzes the current strategies of vocabulary reuse,including use widely-used or standard vocabularies,reuse the same vocabularies,the maximization and minimization of vocabulary number.Finally discusses the characteristics of each strategy.