Research On The Technology Of Schema Matching In Model Management

Publishing view is the primary measure of data exchange. However, themaintenance for publishing view still uses the manual operations basically,which are used to find the changes from the changed source schema, and thenthe experts decide how to deal with the changes. As a result, the automatizationof maintenance for publishing view is one of the issues in the research of modelmanagement, and schema matching is the primary step in the automatic processof maintenance for publishing view. In this paper, based on the syntheticalanalysis to the technologies of schema matching in model management, twonew schema matching algorithms are proposed, and are applied in theprototype, which is designed to deal with the maintenance for publishing XMLview, as follows:Firstly, schema matching algorithm named Similarity Flooding (SF) isanalyzed, the shortcomings and limitations are pointed out, and then animproved SF schema matching algorithm named ExtendMatch is proposed.Compared with SF algorithm, ExtendMatch can be applied in much widerareas.Secondly, the algorithm based on schema matching with opaque columnnames and data values is researched, the operation process is analyzed, and theoptimization and the improvement are proposed to solve the large scale of thecomputation of the algorithm. Compared with the previous algorithm, theimproved algorithm simplifies the computing process, reduces the timecomplexity, and improves the performance of the algorithm.Thirdly, depending on the idea of the schema matching algorithm withopaque column names and data values, a new schema matching algorithmnamed OTTSM(Opaque to Transparent Schema Matching) is proposed, whichcan transform opaque column names and data values to the transparent datavalues using mutual information, and do schema matching based on interzonedistribution of the mutual information. In addition, the application that OTTSMis applied in information security of database is discussed.Finally, the algorithms in the paper are validated by experiments. Basedon the algorithms, a scenario of the prototype is designed, which aims atresolving maintenance for publishing XML view.