9.3 Record Comparator

A RecordComparator is constructed using a list of field comparators. Once field comparison functions have been initialised, they can be inserted into a field_comparisons list which is then given to a record comparator. When comparing records in a linkage or deduplication process, each field comparator in a field_comparisons list will calculate a weight, and all weights for a record pair will be stored in a weight vector. Additional information stored with a weight vector are the unique record identifiers for the two records in the pair compared. The weight vectors will then be used to compute the matching status of the record pair in the classifier as described in Section 9.5.

The RecordComparator is initialised using references to two data sets to be linked (which can be the same in case of a deduplication task), and a list of field comparison functions. An example initialisation is given in the following section.