Research And Implementation On RDF Storage Based On NoSQL And Redundancy Elimination

With the rapid development of the Semantic Web, data in RDF format is also increasing. But nowRDF data storage management system has been unable to adapt to the explosive growth of the data,especially the management system meets the large data storage, high query performance and lowincrease modification etc. When in the open environment, the construction and maintenance ofontology system will appear ambiguity problem. First of all the ontology instances of differentorganizations and personnel for a domain is maintained and added, which may appear repeatedlydescribe instances of the problem and can appear to describe the different instances of the same object,even there are conflicting, polysemy; then when combining with related areas of two and more thantwo ontology, it can appear the same ontology concept but having two ontology description, justdifferent attributs or name. It also calls heterogeneous ontology or ambiguity. At the same time thatthe integration of the user there will be having special preferences and demand for some attribute.How to solve the above problems, this paper will conduct relevant research and implement thecorresponding.In this paper,we have designed and implemented the RDF storage system based on MongDB,which is a document oriented NoSQL database, to manage RDF data. In the design, we have usedappropriate index to improve the query efficiency and have experimented the superiority of thescheme in the performance. In a large-scale ontology, which has the massive instances, if that usingexisting research methods to find the redundant instances, it is not feasible commonly. So we need theprevious pretreatment, and then we use the method of weighted instance of attributes and their valuesto find the redundants. Combined with related areas of two and more ontologies, in consideration ofthe user preferences and needs, we have used a method of eliminating redundant body WeightedConcept Lattice Based on entropy and deviation to meet the above requirements. Besides that we havagiven the weight(B) quantification formula through the experiments. We have verified the methods isfeasible to solve the problems above.