NICTA and UBC at the TREC 2012 Medical Track

Abstract

We introduce two heterogeneous query expansion techniques, and a combined system to the TREC 2012 Medical Track. Our methods are based on external resources that provide expansion concepts related to the query terms, by means of the PageRank algorithm, and simple rules based on UMLS Semantic Types. In this paper we show that our systems are able to reach competitive performances at both the TREC-2011 and TREC-2012 tasks.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Nov 01, 2012
Accession Number
ADA581269

Entities

People

  • Arantxa Otegi
  • David Martinez
  • Eneko Agirre

Organizations

  • University of Melbourne

Tags

DTIC Thesaurus Topics

  • Abstracts
  • Algorithms
  • Automatic
  • Diseases And Disorders
  • Equations
  • Information Operations
  • Language
  • Mathematics
  • Natural Language Processing
  • Natural Languages
  • Pain
  • Pneumonia
  • Probability
  • Probability Distributions
  • Random Walk
  • Standards
  • Ventilators

Fields of Study

  • Computer science

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Clinical Trial Research.
  • Information Retrieval