There are many situations that require the reconciliation of two sets of data. Differences between the sets may be real or insubstantial, and it is the substantial differences that are better surfaced to a user. This disclosure describes techniques to automatically reconcile two sets of data. The techniques are particularly useful when the overlap between the two data sets is unknown or when the elements within the sets appear in different forms. Mismatches between the sets are captured using abstract costs, which enables partial matching between the two sets unconstrained to be one-to-one, and which can happen at multiple levels of granularity or abstraction. The techniques enable a user to customize the tradeoff between false positive and false negative matches based on evolving business objectives.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.