Validation of Matching

TL;DR

This paper presents a method to compute PAC bounds on precision and recall for matching algorithms, useful for network reconciliation and entity resolution, requiring only some verified matches without knowledge of network generation.

Contribution

It introduces a technique to derive PAC bounds on matching algorithm performance using verified matches, applicable to network reconciliation and entity resolution.

Findings

Provides PAC bounds without needing network generation knowledge

Applicable to network reconciliation and entity resolution

Requires only some verified matches for bounds computation

Abstract

We introduce a technique to compute probably approximately correct (PAC) bounds on precision and recall for matching algorithms. The bounds require some verified matches, but those matches may be used to develop the algorithms. The bounds can be applied to network reconciliation or entity resolution algorithms, which identify nodes in different networks or values in a data set that correspond to the same entity. For network reconciliation, the bounds do not require knowledge of the network generation process.