Principles of Data Integration

7

Data Matching

Data matching is the problem of finding structured data items that describe the same real-world entity. In contrast to string matching (see Chapter 4), where we tried to decide whether a pair of strings refer to the same real-world entity, here the entity may be represented by a tuple in a database, an XML element, or a set of RDF triples. For example, we would like to determine whether the tuples (David Smith, 608-245-4367, Madison WI) and (D. M. Smith, 245-4367, Madison WI) refer to the same person.

The problem of data matching arises in many integration situations. In the simplest case, we may be merging multiple databases with identical schemas, but without a unique global ID, and want to decide which rows are duplicates. ...