Research On Hybrid Mapping Of Linked Dataset

As a kind of current network expansion form, semantic network is to build the data that a machine can understand and read and can through data that are rich in semantic annotation to create links between each other, making the whole network into a common medium of exchange of information.For the goal of the semantic web, linked data provides a feasible practical solutions.But as datasets on the network are increasing, the potential of link between datasets are not get a good play.The reason is mainly due to different datasets are using different vocabulary to describe the same entity, it caused the various datasets on the description format is difficult to reach an agreement and links between datasets have obstacles.So in order to play the maximum value of linked data, to achieve real sense of open, smart and share, it must solve the mapping problem between linked datasets.In view of this,this paper presents the hybrid mapping of linked dataset and mainly does the following works:(1) To review and summarize linked dataset mapping problem and the existing technology. First,this paper introduces the related theory and technology of linked dataset,and expounds the problem of mapping, mutual chain situation and existing mapping technologies and their effects of linked dataset.(2) Put forward the Linked datasets hybrid mapping model. Analyzing the network environment,including the vocabulary usage of dataset, metadata description, RDF link conditions and so on. Based on the above, the paper presents the hybrid mapping model of linked dataset and explains the relationship between various levels in hybrid mapping.(3) Expounds the linked dataset hybrid mapping strategy. From the VOID languages metadata descriptions, deployment, and descriptions found to expound the level mapping strategy, and introduced the created and published of the R2R language mapping in detail, finally introduced the WordNet semantic dictionary to computation the semantic similarity.(4) Analyzes the linked dataset hybrid mapping example. With DBpedia. Geonames and LinkedMDB for example, these linked datasets are described by VOID language and to create the R2R mapping between the classes and properties, some specific instances of semantic similarity was calculated, and finally the mapping effect was evaluated.The paper hope that through the use of hybrid mapping of linked dataset to solve the parts of present problems in the dataset mapping, and improve the efficiency of the mapping and mapping accuracy and provide support to create linked data integration service environment.At the same time the research for the linked dataset of mapping method of study provides a new way of thinking, so as to better promote the development of linked data network. MASTER’S THESIS...