Latent argumentative pruning for compact MEDLINE indexing

Authors:
Patrick Ruch;Robert Baud;Johann Marty;Antoine Geissbühler;Imad Tbahriti;Anne-Lise Veuthey
Affiliations:
Medical Informatics Service and Swiss-Prot Group, University Hospital of Geneva and Swiss Institute of Bioinformatics, Geneva, Switzerland;Medical Informatics Service and Swiss-Prot Group, University Hospital of Geneva and Swiss Institute of Bioinformatics, Geneva, Switzerland;Medical Informatics Service and Swiss-Prot Group, University Hospital of Geneva and Swiss Institute of Bioinformatics, Geneva, Switzerland;Medical Informatics Service and Swiss-Prot Group, University Hospital of Geneva and Swiss Institute of Bioinformatics, Geneva, Switzerland;Medical Informatics Service and Swiss-Prot Group, University Hospital of Geneva and Swiss Institute of Bioinformatics, Geneva, Switzerland;Medical Informatics Service and Swiss-Prot Group, University Hospital of Geneva and Swiss Institute of Bioinformatics, Geneva, Switzerland
Venue:
AIME'05 Proceedings of the 10th conference on Artificial Intelligence in Medicine
Year:
2005

Citing 6
Cited 0

Pivoted document length normalization

SIGIR '96 Proceedings of the 19th annual international ACM SIGIR conference on Research and development in information retrieval
Extended Boolean information retrieval

Communications of the ACM
Static index pruning for information retrieval systems

Proceedings of the 24th annual international ACM SIGIR conference on Research and development in information retrieval
Probabilistic models of information retrieval based on measuring the divergence from randomness

ACM Transactions on Information Systems (TOIS)
Using contextual spelling correction to improve retrieval effectiveness in degraded text collections

COLING '02 Proceedings of the 19th international conference on Computational linguistics - Volume 1
Distribution of information in biomedical abstracts and full-text publications

Bioinformatics

Quantified Score

Hi-index	0.00

Visualization

Abstract

PURPOSE: We evaluate how argumentation in scientific articles can be used to propose an original index pruning strategy, which significantly reduce the size of the engine's indexes but having a limited impact on retrieval effectiveness. METHODS: A Bayesian classifier trained on explicitly structured MEDLINE abstracts generates these argumentative categories. The categories are used to generate four different argumentative indexes. A fifth index contains the complete abstract, together with the title and the list of Medical Subject Headings (MeSH) terms. This last index is used as baseline to compare results obtained when only a specific argumentative index is retrieved. RESULTS and CONCLUSION: When titles and medical subject headings are also stored in the respective indexes, querying PURPOSE and CONCLUSION indexes can respectively achieves 78.4% and 74.3% of the baseline, while the size if the index is divided by two. It is concluded that argumentation can be a powerful index pruning strategy in complement to more traditionnal approaches.