Blocking: ChoiceMaker software selects candidate records that might match each other
Scoring: For each pair of candidate records, ChoiceMaker computes the probability that the records match each other, and recommends a decision for how to process the pair.
Match: Automatically process that records as representing the same person
Differ: Automatically process the records as representing different persons
Hold: Set the pair of records aside for human review by data experts
Transitivity analysis: If more than two records potentially match each other, ChoiceMaker gathers the records together as a group and analyzes whether the group displays any internal clustering.
ClueMaker is an easy-to-learn scripting language that intuitively expresses common types of field comparisons. With ClueMaker, simple comparisons are easy and complex comparisons are possible. For example, a few lines of code can specify the exact or fuzzy comparison of a multi-valued field between two records, with full validation of each field value.
ChoiceMaker libraries provide numerous algorithms for both exact and fuzzy comparisons of record fields. ChoiceMaker libraries include text-matching routines for Soundex, Edit-distance, Jaro-Winkler, NYSIIS, Metaphone, Double-Metaphone, and value-frequency weighting. Third-party libraries can be incorporated as plugins to provide additional capability.
ChoiceMaker software is designed to be extensible. It is built on a plugin framework that makes it easy to customize the software for specific client requirements or to add new capabilities to the core engine. ChoiceMaker currently has several extensions under development and ChoiceMaker welcomes extensions by third-party developers.
ChoiceMaker products and extensions are available through various packages. All packages provide software support and access to the latest source and binary code. In addition, ChoiceMaker software is released annually as open source under the Eclipse Public License.
Learn more with the ChoiceMaker Technical Overview