Incorporating string transformations in record matching

Today's record matching infrastructure does not allow a flexible way to account for synonyms such as "Robert" and "Bob" which refer to the same name, and more general forms of string transformations such as abbreviations. We expand the problem of record matching to take such user-defined string transformations as input. These transformations coupled with an underlying similarity function are used to define the similarity between two strings. We demonstrate the effectiveness of this approach via a fuzzy match operation that is used to lookup an input record against a table of records, where we have an additional table of transformations as input. We demonstrate an improvement in record matching quality and efficient retrieval based on our index structure that is cognizant of transformations. Categories and Subject Descriptors H.2.4 [Database Management]: Systems; H.2.8 [Database Management]: Database Applications General Terms Algorithms, Performance, Experimen...

• Type: Journal
• Year: 2008
• Where: SIGMOD
• Authors: Arvind Arasu, Surajit Chaudhuri, Kris Ganjam, Raghav Kaushik
• Keywords: Database, Database Management, Fuzzy Match Operation, SIGMOD 2008, User-defined String Transformations