Artificial Intelligence in Medicine
Volume 51, Issue 1 , Pages 17-25 , January 2011

Exploiting the systematic review protocol for classification of medical abstracts

  • Oana Frunza

      Affiliations

    • School of Information Technology and Engineering, University of Ottawa, 800 King Edward, Ottawa, Ontario, Canada K1N 6N5
    • Corresponding Author InformationCorresponding author. Tel.: +1 613 562 5800x2140; fax: +1 613 562 5175.
    • ,
  • Diana Inkpen

      Affiliations

    • School of Information Technology and Engineering, University of Ottawa, 800 King Edward, Ottawa, Ontario, Canada K1N 6N5
    • ,
  • Stan Matwin

      Affiliations

    • School of Information Technology and Engineering, University of Ottawa, 800 King Edward, Ottawa, Ontario, Canada K1N 6N5
    • ,
  • William Klement

      Affiliations

    • School of Information Technology and Engineering, University of Ottawa, 800 King Edward, Ottawa, Ontario, Canada K1N 6N5
    • ,
  • Peter O’Blenis

      Affiliations

    • Evidence Partners Corporation, 9 Wick Crescent, Ottawa, Ontario, Canada K1J 7H1

Received 18 January 2008 ,Revised 22 September 2010 ,Accepted 14 October 2010.

  • Image Result

    Embedding automatic text classification in the process of building a systematic review.

    Embedding automatic text classification in the process of building a systematic review.

  • Image Result

    Example of confusion matrix [TI=the number of true inclusions; FE=the number of false exclusions; FI=the number of false inclusions; TE=the number of true exclusions].

    Example of confusion matrix [TI=the number of true inclusions; FE=the number of false exclusions; FI=the number of false inclusions; TE=the number of true exclusions].

  • Image Result

    Algorithm for per-question classification method. Recall and precision plots when varying the training size for per-question technique.

    Algorithm for per-question classification method. Recall and precision plots when varying the training size for per-question technique.

  • Image Result

    Summary of results for both the global and per-question method (for the per-question method, the voting scheme used is indicated in the names by subscript numbers).

    Summary of results for both the global and per-question method (for the per-question method, the voting scheme used is indicated in the names by subscript numbers).

PII: S0933-3657(10)00124-7

doi: 10.1016/j.artmed.2010.10.005

Artificial Intelligence in Medicine
Volume 51, Issue 1 , Pages 17-25 , January 2011

Article Tools

* Image Usage & Resolution