Data File Builder

Introduction

The MetaMap application is designed to automatically identify UMLS Metathesaurus concepts referred to in free text. Although the UMLS focuses on biomedical information sources, MetaMap's algorithms are domain independent and can be used with any domain given adequate knowledge sources. The MetaMap Data File Builder enables such cross-domain utilization of MetaMap by allowing users to create UMLS-like data models similar to the actual UMLS data models normally used by MetaMap.

Brief documentation on installing the MetaMap Data File Builder is in the Datafile Builder README. Instructions on how to use datafile builder is in the MetaMap Data File Builder Manual. Mac OS/X users see also the file README_macosx.html (online version)

Addendum to Datafile Builder Manual: In early versions of the 2013 Manual, Section 7.2 "Using a UMLS Metathesaurus Subset" incorrectly specified "Original Release Format" for the "Select Output Format" box; "Rich Release Format" should be selected.

More on creating a MetaMap Dataset

A short article on creating a MetaMap dataset from the EFO Inferred Ontology is in this document: Transforming the EFO Inferred Ontology for MetaMap (PDF). Both a zip archive (efo2dfb.zip) and a tar-gzipped archive (efo2dfb.tar.gz) containing the source code is available.