Implementing Boolean Matching Rules in an Entity Resolution System using XML Scripts

摘要

This paper describes a method for allowing users to define Boolean matching rules for entity resolution systems using XML scripts. The design allows users the flexibility to create and modify matching rules that are interpreted at run-time rather than having to go through the process of modifying and testing the underlying codebase. The design has been validated through implementation in open source entity resolution system named OYSTER. The paper also describes proposed design enhancement that address the issues of cross-attribute comparison and conflict rules.