OHSUMED: an interactive retrieval evaluation and new large test collection for research

By William R. Hersh, Chris Buckley, T. Leone, David H. Hickam. 1994

Subfields: Information Retrieval & Search, Evaluation & Benchmarks

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Information Retrieval & Search

In Evaluation & Benchmarks

Citations in the corpus

Super-NaturalInstructions: Generalization via Declarative Instructions on 1600+ NLP Tasks (2022, methodology): “Table 13: An example task in the Keyword Tagging category of our dataset, adopted from OHSUMED (Hersh et al., 1994).”
Hierarchical Text Categorization Using Neural Networks (2004, methodology): “More specifically, we explore a categorization strategy designed to exploit the hierarchical structure underlying the UMLS (Unified Medical Language System) Metathesaurus (National Library of Medicine 1999) and test its effectiveness using the OHSUMED test collection (Hersh et al. 1994).”
Introduction to Information Retrieval (2008, unknown): “Hersh et al. (1994) present similar analysis for a medical IR collection. In contrast, Kekäläinen (2005) analyze some of the later TRECs, exploring a 4-way relevance judgment and the notion of cumulative gain, arguing that the relevance measure used does substantially affect system rankings. See als”
Overcoming the Brittleness Bottleneck using Wikipedia: Enhancing Text Categorization with Encyclopedic Knowledge (2006, background): “OHSUMED (Hersh et al. 1994) is a subset of MED- LINE, which contains 348,566 medical documents.”
AdaRank: a boosting algorithm for information retrieval (2007, methodology): “2 Experiment with OHSUMED Data In this experiment, we made use of the OHSUMED dataset [14] to test the performances of AdaRank.”
Statistical topic models for multi-label document classification (2011, methodology): “ the Yahoo! subdirectory datasets have removed the infrequent labels from the evaluation data, leaving between 14 and 23 unique labels per dataset. Similarly, experiments with the OHSUMED MeSH terms (Hersh et al., 1994) are typically performed on a small subdirectory that contains only 119 out of ov”
Multitask Bregman clustering (2010, methodology): “The ohscal data set was obtained from the OHSUMED collections [42], which are also academic paper abstracts on medical.”
Assessing thesaurus-based query expansion using the UMLS Metathesaurus (2000, methodology): “As noted already, it has been shown that the presence of these MeSH terms leads to a 10% improvement in performance as measured by average precision [15].”
Support Vector Ordinal Regression (2007, background): “Hersh et al. (1994) generated the OHSUMED dataset,(8) which consists of 348566 references and 106 queries with their respective ranked results.”
Test Collection Based Evaluation of Information Retrieval Systems (2010, background): “[118] released a test collection composed of around 350,000 catalogue entries for scholarly articles.”
Collective multi-label classification (2005, methodology): “The OHSU-Med [7] HD corpus, a popular dataset for text classi.cation, is a collection of titles and abstracts of medical research journal articles from 1989-1991 corresponding to characterizations of the relevant heart conditions, such as Heart Aneyurism and Myocarditis .”