Research On Rule Based Linked Data Compression

With the standardization of Semantic Web technology,linked data is widely used in various fields,which makes the publishing of linked data to an unprecedented scale.This leads to the difficulty of storage and transmission of linked data.The method to reduce the volume of linked data via compression has become a hot spot in Web research area.The linked data compression method can be divided into two categories,which are syntactic compression and semantic compression,and in this paper we mainly study the semantic compression.We proposed a rule based linked data compression method.This method mined rules from linked and remove the triples which can be inferred by these rules,so as achieve the purpose of compression.Rule mining is the core of the compression method which will affect the compression ratio and compression efficiency.In this paper,we considered the structure of linked data and adopted a rule mining method based on data set block.First,we put the entities with the same entity description pattern together to form a data set block,then we extract rules from each data set block,finally we merge these rules to form the rules of linked data.With the experimental confirmation and result analysis on real datasets,our method is able to remove about 30% of the original triples without affecting the semantic integrity of linked data.And it is better than the existing semantic compression method in the category of rules,compression ratio and compression efficiency.