A Versatile Record Linkage Method by Term Matching Model ...

A Versatile Record Linkage Method by TermMatching Model Using CRFQ uang Minh Vu, Atsuhiro Takasu, and Jun AdachiNational Insitute of Informatics, Tokyo 101-8430, solve the problem of Record Linkage between databaseswhere Record fields are mixed and permuted in different ways. The so-lution Method uses a conditional random fields Model to find matchingterms in Record pairs and uses Matching terms in the duplicate detec-tion process. Although records with permuted fields may have partlyreordered terms, our Method can still utilize local orders of terms forfinding Matching terms. We carried out experiments on several well-known data sets in Record Linkage research, and our Method showed itsadvantages on most of the data sets. We also did experiments on a syn-thetic data set, in which records combined fields in random order, andverified that it could handle even this data IntroductionInformation on the web is growing at an explosive rate [10], and informationabout an object may appear in several places.

A Versatile Record Linkage Method by Term Matching Model Using CRF Quang Minh Vu, Atsuhiro Takasu, and Jun Adachi National Insitute of Informatics, Tokyo 101-8430, Japan

Loading..

Tags:

  Versatile, Methods, Record, Linkages, A versatile record linkage method by

Information

Related search queries