JMatching: A Sample-Driven Schema Matching

Abstract

Schema mappings are high-level specifications that describe the relationship between schemas. In real-life applications schema mappings can be a quite complex task since users must know the semantic of the schemas to be matched. However, there exist applications that users know the domain but not the details of the schemas. Thus, tools that can build a target schema based on samples are demanded by users non-experts in computer science but experts in the application domain. This work proposes an approach for schema mapping based on samples to help non-experts users to build their own database. Some experiments were conducted to verify schemas and built database correcteness. Besides, we measured the amount of memory needed and execution time. In all experiments, our work showed good results.

Keywords

• schema matching
• sample-driven
• data integration
• database