Measuring duplicate metadata records in library databases

摘要

The changing architecture of bibliographic databases has created the problem of duplication of journal article metadata. Metadata has become a commodity, and packages of it are licensed to companies who redistribute it and make it available to users. Duplicate records have long been a problem for monograph and serial bibliographic records, and now with the increase and cross-licensing of journal article metadata, there is a need to consistently measure and define duplication in metadata databases. This article presents a new measure of duplication, which is the proportion of duplicate records retrieved in a search to the total number of records retrieved in the search. This equation measures and represents the concept of duplication in bibliographic databases. This measure can be used to compare duplication among databases and to better focus efforts towards the deduplication of metadata in bibliographic databases.