A Biomedical Information Retrieval System based on Clustering for Mobile Devices

993 views

Published on

Published in: Technology
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
993
On SlideShare
0
From Embeds
0
Number of Embeds
89
Actions
Shares
0
Downloads
17
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

A Biomedical Information Retrieval System based on Clustering for Mobile Devices

  1. 1. A Biomedical Information Retrieval System <br />based on Clustering for Mobile Devices<br />Manuel de la Villa<br />Manuel Millán<br />Alejandro Muñoz<br />Manuel J. Maña<br />This work has been partially funded by the Spanish Ministry of Science and Innovation and the European Union from the ERDF (TIN2009-14057-C03-03)<br />1<br />
  2. 2. Mainindex<br />Introduction<br />Search and InformationRetrieval<br />Clustering<br />Visualizationon Mobile Devices<br />Conclusions and Future Works<br />2<br />
  3. 3. Mainindex<br />Introduction<br />Search and InformationRetrieval<br />Clustering<br />Visualizationon Mobile Devices<br />Conclusions and Future Works<br />3<br />
  4. 4. Introduction<br />Motivation<br />Medicalstaffmobility<br /><ul><li>“Hospitals staff might be distributed in space or time and their information needs are highly dependent on contextual conditions.“ (Muñoz et al, 2003)
  5. 5. “…PDA use by health professionals shows an evolution in the use ranging from 30% in 2000 to 60% in 2006” (Garrity and El Emam, 2006)
  6. 6. the available resources accessible from the PDA at the bedside provided response to 86% of clinical questions, most of them (88.9% - 97.7%) during the rounds of visits (Hauser et al. 2007)</li></ul>4<br />
  7. 7. Introduction<br />Motivation<br />Evidence-based medicine<br /><ul><li>“Evidence based medicine is the conscientious, explicit, and judicious use of current best evidence in making decisions about the care of individual patients. The practice of evidence based medicine means integrating individual clinical expertise with the best available external clinical evidence from systematic Research. “(Sackett et Al., 1996)
  8. 8. Isusefulan IRS tolocatethe best available external clinical evidence?</li></li></ul><li>Introduction<br />Motivation<br />Wi-fi<br />Informationoverload<br />PDA<br />Point –of-care<br />Bestevidence<br />Mobility<br />CLUMMED<br />Efficientaccess<br />CLUsteringonMobile MEdicalDevices<br /><ul><li>YAIRS? YetAnother IRS? Novelty?
  9. 9. Post-retrievalclustering, orientationtobiomedicaldocuments and mobiledevices</li></li></ul><li>Introduction<br />Post-retrievalclustering?<br /> Post-retrievalclusteringis a knowntecniquethatimprovetheorganization of thesearchresults and facilitatenavigationbetweenthem.<br />PreviousexperiencesonSearchEnginesusingclustering:<br /> Univesity Carnegie-Mellon -> Vivisimo -> Clusty<br />Interesting? No?<br />And if I tellyouthatonMayYippy has paid$5.5M forClusty…?<br />7<br />
  10. 10. Introduction<br />Post-retrievalclusteringonBiomedicine?<br />
  11. 11. Index<br />Introduction<br />Search and InformationRetrieval<br />Clustering<br />Visualizationon Mobile Devices<br />Conclusions and Future Works<br />9<br />
  12. 12. Search and InformationRetrieval<br />A tipicalSchema<br />InformationRetrievalSystem<br />Given a set of documents and an information need, the goal of IR is to obtain the documents relevant to that need, sort by any criteria and show them to the user.<br />Documentsrepresentation<br />Documents<br />Repository<br />Indexing<br />Query<br />Searching<br />Evaluation<br />Ranking<br />Textprocessing<br />Similaritycalculation<br />Informationneeded<br />Queryrepresentation<br />Analysis<br />RelevantDocuments<br />
  13. 13. Search and InformationRetrieval<br />Ourimplementation<br />Documentsources:Biomed Central (web crawling in progress)<br />TextProcessing:lowercasing, stemming, stop-words ,…<br />Lucene for indexing…<br />
  14. 14. Search and InformationRetrieval<br />Ourimplementation (and II)<br />… and Lucene for searching<br />
  15. 15. Index<br />Introduction<br />Search and InformationRetrieval<br />Clustering<br />Visualizationon Mobile Devices<br />Conclusions and Future Works<br />13<br />
  16. 16. Clustering<br />Ourimplementation<br />Clustering<br />The post-processing clustering is to associate, according to their similarity, a set of documents retrieved from a query in different subsets<br />14<br />
  17. 17. Clustering<br />Why Simple-K-Means?<br />Clusteringalgorithm:<br />Simple-K-Means vs ExpectationMaximization<br />Time it takes to perform the grouping in seconds<br />K? Itdependsonthenumber of documentsretrieved.<br />15<br />
  18. 18. Index<br />Introduction<br />Search and InformationRetrieval<br />Clustering<br />Visualizationon Mobile Devices<br />Conclusions and Future Works<br />16<br />
  19. 19. Visualizationon Mobile Devices<br />Typicalproblems<br />Visualizationon Mobile Devices<br /><ul><li>Restrictions:
  20. 20. Compatibilityproblems
  21. 21. Screensize
  22. 22. Multipleswindowslack
  23. 23. Limitednavigation
  24. 24. Limitedmemorysize
  25. 25. Javascript, cookies
  26. 26. Accesibility
  27. 27. etc.
  28. 28. Solutions:
  29. 29. World Wide Web Consortium (W3C): Mobile Web Initiative. </li></ul>"The Mobile Web Initiative’sgoalistomakebrowsingthe Web frommobiledevices a reality, toimprove Web contentproductionandaccessformobileusers” (Tim Berners-Lee, W3C Director) <br /><ul><li>Mobile Web BestPractice 1.0.</li></ul>17<br />
  30. 30. Visualizationon Mobile Devices<br />Some best practices<br />One web<br />making, as far as is reasonable, the same information and <br />services available to users irrespective of the device they are using<br />Trust in web standards<br /> HTML compatible with different browsers, Use Stylesheet, <br />Content in blocks (<DIV>).<br />Avoid known risks<br />No pop-ups, No frames, No tables<br />Controlling limitations<br />No scripting, Standards fonts, Use of color<br />Optimized navigation<br />Minimal navigation at the top of the page<br />Avoid lengthy URI’s<br />Probe images and colours<br />Reduced image size and resolution, good contrast<br />Do it small<br />Small pages, only one-direction (vertical) scrolling, easy entry forms (reduced keytyping)<br />Limited use of network<br />No external links (images…), no download<br />Think in users<br />Simple language, relevant and limited content, error messages<br /> And many more…!!!<br />18<br />
  31. 31. Visualizationon Mobile Devices<br />Highresolutioninterface<br />19<br />
  32. 32. Visualizationon Mobile Devices<br />Low resolutioninterface<br />20<br />
  33. 33. Visualizationon Mobile Devices<br />CheckingClusterMed<br />21<br />
  34. 34. Visualizationon Mobile Devices<br />Checkingourproposal<br />Ourproposalobteins at hisfirstversion a 76% <br />
  35. 35. Visualizationon Mobile Devices<br />Ourinterface<br />Cancerskin<br />23<br />
  36. 36. Introduction<br />Search and InformationRetrieval<br />Clustering<br />Visualizationon Mobile Devices<br />Conclusions and Future Works<br />24<br />
  37. 37. Conclusionsandfuturework<br />Conclusions<br />Conclusions<br /><ul><li>ObjectivesFulfilled
  38. 38. Itworks!!!
  39. 39. FirstMilestone
  40. 40. A workingprototype of anInformationretrievalSystemadaptedforMedicalDevicesbasedon Post-retrievalclustering</li></ul>25<br />
  41. 41. Conclusionsandfuturework<br />Futurework<br />… and future work:<br /><ul><li>A lot of…
  42. 42. FocusonusingmedicalontologieslikeUMLS Metathesarusfor:
  43. 43. Improvetheclusteringquality
  44. 44. Enhance the labelling of the groups
  45. 45. Visual help, graph with concepts/cluster relationship</li></ul>26<br />
  46. 46. Conclusionsandfuturework<br />Futurework (and II)<br />… and future work (and II):<br /><ul><li>Monodocument Summarization
  47. 47. Freebase Ajax connection for search support
  48. 48. Bilingual (Spanish-English documents)
  49. 49. Put into production (crawling and indexing…) any sponsor at the hall???</li></ul>27<br />
  50. 50. A Biomedical Information Retrieval System <br />based on Clustering for Mobile Devices<br />Manuel de la Villa<br />Manuel J. Maña<br />{manuel.villa, manuel.mana}@dti.uhu.es<br />Manuel Millán<br />Alejandro Muñoz<br />{manuel.millan, alejandro.munoz}@alu.uhu.es<br />This work has been partially funded by the Spanish Ministry of Science and Innovation and the European Union from the ERDF (TIN2009-14057-C03-03)<br />28<br />

×