OHSUMED - sporedata/researchdesigneR GitHub Wiki

General description

The OHSUMED dataset comprises 348,566 references from MEDLINE, consisting of abstracts and/or titles from 270 medical journals over five years (1987-1991). The available fields are abstract, author, MeSH indexing terms, publication type, source, and title.

The National Library of Medicine (NLM) has permitted access to the MEDLINE references in the test database for experimentation under two conditions:

  1. The data will not be used in any non-experimental clinical, library, or other setting.
  2. All users of the data will be expressly informed that the data is incomplete and out-of-date.

Related publications

OHSUMED: An Interactive Retrieval Evaluation and New Large Test Collection for Research

Data access

OHSUMED Dataset