Please use this identifier to cite or link to this item:
Title: Record matching in digital library metadata
Authors: Kan, M.-Y. 
Tan, Y.F.
Issue Date: 2008
Citation: Kan, M.-Y., Tan, Y.F. (2008). Record matching in digital library metadata. Communications of the ACM 51 (2) : 91-94. ScholarBank@NUS Repository.
Abstract: Record matching in digital library metadata uses evidences from the external sources to create more accurate matching systems. A manually compiled set of rules named a crosswalk may be used to transform the metadata into the library's own format, when libraries import large volumes of metadata from sources that follow a metadata standard. Graphical formalisms in the form of generative probabilistic models have also been suggested in which the model picks out collaborative groups and then assigns authors within this groups to generate references. The model can then run this model in the opposite direction to infer which collaborative group is responsible for a particular work. The graphical models have outperformed methods using pairwise comparisons in accuracy.
Source Title: Communications of the ACM
ISSN: 00010782
DOI: 10.1145/1314215.1340938
Appears in Collections:Staff Publications

Show full item record
Files in This Item:
There are no files associated with this item.

Google ScholarTM



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.