Semantic interoperability often relates to linking files based upon probabilistic matching of files. Matching files has a long history of uses and with the advances in technology, it has become technically practical to follow the health experience and mortality of large populations who are thought to be as some special risk by linking existing health and death records for individuals across multiple databases. The mayor categories of record linking are "deterministic" or "exact" matching and "probabilistic". Generally, such linkages must be probabilistic if the are to be reasonably complete.
In his "Handbook of Record Linkage published in 1959, Howard B. Newcomb, Ph.D. defines linkage as: "Record linkage is the joining of information from two records that are believed to relate to the same individual, family, event, business or address (geo-coding). The slides below depict two basic record sets that were matched through iterative processing on birth weight, data of birth, hospital, zip code and county. An approximate match of 95% was realized at the first stage while the remainder were probabilistically matched. The resulting data set was constructed for public use.
In his "Handbook of Record Linkage published in 1959, Howard B. Newcomb, Ph.D. defines linkage as: "Record linkage is the joining of information from two records that are believed to relate to the same individual, family, event, business or address (geo-coding). The slides below depict two basic record sets that were matched through iterative processing on birth weight, data of birth, hospital, zip code and county. An approximate match of 95% was realized at the first stage while the remainder were probabilistically matched. The resulting data set was constructed for public use.
(to be continued)
No comments:
Post a Comment