Research On Method Of Discovering Relationship Among Data Sources In Dataspace

Today’s society, rapid growth data emerges mass characteristics，diversity, heterogeneity and magnanimity, A variety of data managementsystems are available. These management methods are model-drivenmanagement style basically. Model-driven management approach can’t solvethe bottlenecks and challenges that encountered in the data management.DataSpace is raised in such a situation. The DataSpace is a collection of alldata and their relationships between data and the user. It supports a variety ofheterogeneous data forms, has the ability to self-evolution, adopts to thepay-as-you-go evolution way, Best-effort query service mode, dynamicallyextract data mode. It provides convenient help for users to store, query, search,update, and manage data. In order to achieve unified management of multipledata sources and achieve the way of lightweight, efficient data management,This thesis is mainly to study a way of discovering the correlation between theheterogeneous data sources in DataSpace. The main tasks are summarized asfollows:First, This article first proposed the word relation model, it is the basic offinding the association between data source content. Through analyzing theHowNet’s hierarchy organizational form and concept, the article use HowNetcorpus to compute the words relation. The word relation model can be used tocalculate the relation of the same kinds of parts of speech, as well as thedifferent parts of speech. It integrates word similarity degree, wordassociation degree, and examples of factors, Taken together as the wordsrelation. Combination with experiment verification, we find the words relationmodel that is proposed in this paper, can calculate words relation degree andthe results are more subjective understanding of the human.Second, From the point of view of natural language processing, based on theword relation model, this article designs a discovery mechanism of data source content in DataSpace. It provides a good basis for the research group to create anindex, browse, search, query, and other services.The paper researches the method of discovering the association mechanism ofDataSpace content, based on the word relation model, from the angle of the facetswhich reflect the data content. do some exploratory work for further study for ourresearch group.