Font Size: a A A

Research On Link Traversal-Based Query Execution Of Linked Data

Posted on:2015-02-28Degree:MasterType:Thesis
Country:ChinaCandidate:J C HouFull Text:PDF
GTID:2308330482460331Subject:Computer system architecture
Abstract/Summary:PDF Full Text Request
During recent years an increasing number of linked data is published on the Web, thus creating a globally distributed dataspace.While due to the distributed and opennessnature of this dataspace, query execution on the Web of Data faces various challenges. Currently, how to effectively query and use the linked data has become a research hot spot.Different from the traditional query methods which replicate remote data into local warehouses and execute query on it, the LTBQE introduce a new query execution mode:it does not rely on the prior knowledge of the related data source but find the related data source in the execution time. Moreover, it can ensure the query result up-to-date and it exploiting the Web of Data to its full potential.In order to improve the query efficiency, this thesis forwards a caching strategy. Firstlly, it puts forward an index to cache data. Here to improve the efficiency of the cache and save the finite memory, it only caches those frequent query patterns and their relevant data sources and it does not cache the concrete data. Secondly, according to the query log, it mines the frequent query patterns and through data source selection algorithm it gets frequent query patterns’ relevant data sources and then cache them into memory. When query the frequent query patterns again that exists in the cache, according to the index the query engine can directly get the related data sources, thus the query time is reduced.Secondly, In order to use the semantic information of linked data and improve the recall of query, this article proposes the reasoning mechanism of LTBQE. About the reasoning mechanism of LTBQE, the thesis mainly does the follow works. Firstly, it extends the inference rules on the basis of the original research to get more results through inference. Then it describes the process of reasoning and put forwards the reasoning algorithm, thus applies terminological data to the inference over LTBQE. Finally, extensive experiments on real datasets show that the algorithm performs well. Meanwhile, it improves the efficient of the queriesby the data source selection and increases the results of the queries through inference.
Keywords/Search Tags:Linked Data, Link Traversal, SPARQL, Reasoning, Source Selection
PDF Full Text Request
Related items