Research On Hierarchical Tree-Index Based Linked Data Compression

With the development and application of semantic web,RDF dataset with large scale was using more and more frequently.While manage these large scale datasets,due to the size problem of these files,the performance of the query operation will be greatly affected.Therefore,for large scale RDF data,efficient storage method is clearly very important.For data compression storage,it is essentially the removal of redundant information contained in the dataset.At present,the compression methods of large scale RDF data can be divided into the following two types: compression processing based on data model level and compression processing based on serialization level,and compression processing based on serialization level can be subdivided into syntactic compression and symbolic compression.In this thesis,we based on the serialization level of RDF data,by analyzing the current technology of RDF data compression,a compression model based on hierarchical treeindex idea and the three-dimensional matrice structure has been proposed.By analyzing and extracting the relationship between the URI of RDF to build the dictionary to map triples into ID triples.By analyzing the three-dimensional matrice structure of linked data,combine the idea that makes hierarchical tree-index to solve the redundant problem of matrice structure.In addition,this thesis also analyzes and summarizes the rules and performance of query on this model.With the experimental confirmation and result analysis,the compression model proposed can effectively improve the compression ratio of RDF data,and it also improves the speed of node searching.