Mining Hierarchies and Similarity Clusters from Value Set Repositories

Kevin J. Peterson, Guoqian Jiang, Scott M. Brue, Feichen Shen, Hongfang Liu

Research output: Contribution to journalArticlepeer-review

2 Scopus citations


A value set is a collection of permissible values used to describe a specific conceptual domain for a given purpose. By helping to establish a shared semantic understanding across use cases, these artifacts are important enablers of interoperability and data standardization. As the size of repositories cataloging these value sets expand, knowledge management challenges become more pronounced. Specifically, discovering value sets applicable to a given use case may be challenging in a large repository. In this study, we describe methods to extract implicit relationships between value sets, and utilize these relationships to overlay organizational structure onto value set repositories. We successfully extract two different structurings, hierarchy and clustering, and show how tooling can leverage these structures to enable more effective value set discovery.

Original languageEnglish (US)
Pages (from-to)1372-1381
Number of pages10
JournalAMIA ... Annual Symposium proceedings. AMIA Symposium
StatePublished - 2017

ASJC Scopus subject areas

  • Medicine(all)


Dive into the research topics of 'Mining Hierarchies and Similarity Clusters from Value Set Repositories'. Together they form a unique fingerprint.

Cite this