Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Free-text medical document retrieval via phrase-based vector space model.

Wenlei Mao1, Wesley W Chu

  • 1Computer Science Department, University of California, Los Angeles, CA, USA.

Proceedings. AMIA Symposium
|December 5, 2002
PubMed
Summary

This study introduces a phrase-based vector space model (VSM) to enhance information retrieval accuracy. The new model significantly outperforms traditional stem-based approaches by incorporating conceptual similarity and common word stems.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Cognitive ontologies for neuropsychiatric phenomics research.

Cognitive neuropsychiatry·2009
Same author

IndexFinder: a method of extracting key concepts from clinical texts for indexing.

AMIA ... Annual Symposium proceedings. AMIA Symposium·2004
Same author

Modeling medical content for automated summarization.

Annals of the New York Academy of Sciences·2003
See all related articles

Area of Science:

  • Information Science
  • Computer Science
  • Medical Informatics

Background:

  • Vector Space Models (VSM) are foundational for information retrieval, representing documents as index term vectors.
  • Concept-based indexing aims to improve retrieval accuracy over traditional word stems.
  • Previous concept-based systems have not consistently outperformed stem-based methods due to knowledge source limitations.

Purpose of the Study:

  • To address the limitations of concept-based indexing by proposing a novel phrase-based representation for documents.
  • To improve information retrieval accuracy by leveraging both conceptual similarity and shared word stems within phrases.
  • To evaluate the effectiveness of the phrase-based VSM against traditional stem-based models.

Main Methods:

  • Documents are represented using phrases, which are combinations of concepts and word stems.

Related Experiment Videos

  • Phrase similarity is calculated based on shared conceptual similarity and common word stems.
  • Document similarity is aggregated from individual phrase similarities.
  • Experiments were conducted using the OHSUMED test collection and UMLS knowledge source.

    • Main Results:

    • The proposed phrase-based VSM achieved a 16% increase in retrieval accuracy compared to the stem-based model.
    • This improvement demonstrates the efficacy of incorporating phrase-level information.
    • The joint determination of phrase similarity proved effective in enhancing retrieval performance.

    Conclusions:

    • Phrase-based representation offers a significant advancement over traditional stem-based methods in information retrieval.
    • Leveraging both conceptual and lexical information within phrases enhances retrieval accuracy.
    • The findings suggest that phrase-based VSM is a promising approach for improving search system performance.