Research On Pattern Discovery And Storage Of XML Data

As a kind of simple, open and extensible self-describing language, XML is not only an emerging Internet standard for data representation and exchange, but also applied to Internet search engines, electronic commerce, EDI, etc. So such attention has been paid to XML. How to extract schema from XML documents and store XML data is an essential problem in XML-related research, which directly affects the efficiency and function of XML documents.In this paper, we discuss the meaning of frequent pattern mining tasks, and introduce a series of frequent pattern mining algorithms. Based on the existing frequent tree structure mining algorithm—WL, we present the XML data model XOEM and compressed structure tree, improve the pruning strategies of candidates and then bring forward an effective algorithm for XML frequent pattern discovery—XFPD. After that, we further research on the formalization of frequent patterns produced by XFPD, which transforms XML frequent patterns to corresponding DTDs. The mapping and optimizing strategies for the generation of relational schema from DTD are introduced to store frequent pattern data. The problem of how to store overflow data out of frequent patterns is also discussed. Using the relational schemas generated as above, we present the automatic importing method of XML data and implement efficient storage of XML data in relational databases.This paper proves that XFPD has better performance than WL in the work of mining XML frequent patterns theoretically and experimentally, and gives a proof of the validity and feasibility of the XML data storage method presented in this paper by the implementation of an antetype model.