GoPubMed versus PubReMiner   for analyzing PubMed search results: A head to head comparison of two free web ‘data mining’ ...
Introduction  Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011 <ul><li>Searching PubMed/...
<ul><li>PubReMiner  is a data mining tool, mining  PubMed - MEDLINE  abstracts  </li></ul><ul><li>Query entered as text, u...
<ul><li>GoPubMed  - a  knowledge-based semantic browsing tool  for life sciences </li></ul><ul><li>One of the first  web 2...
Methods  Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011 <ul><li>A series of PubMed sea...
Results <ul><li>Publication year </li></ul><ul><li>PubReMiner (PRM) displays chronologically </li></ul><ul><li>GoPubMed (G...
Author Names   <ul><li>Giustini D*[au] entered in PubMed:  8 records found – PMIDs entered </li></ul><ul><li>PubReMiner  f...
Country  <ul><li>PRM: </li></ul>PRM: vs GPM: Minor discrepancies found with both tools: a ‘tie’ Sue Bradley, Dean Giustini...
Journal names <ul><li>GoPubMed enters different journal abbreviations for same journal  </li></ul><ul><li>eISSN/ISSN given...
MeSH terms <ul><li>Presented very differently in PubReminer & GoPubMed </li></ul><ul><li>PRM includes MeSH subheadings: </...
<ul><li>GoPubMed presents a neater table; no subheadings </li></ul><ul><li>But you may need to  click through   many, many...
PubReminer  vs Actual PubMed Results MeSH terms: Accuracy Two indexed records  were not included : PMIDs:  20579633 / 2010...
<ul><li>MeSH terms & chemical substance names NOT  read  for two PMIDs:  20579633  20107104   </li></ul>Sue Bradley, Dean ...
GoPubMed  vs Actual PubMed Results MeSH terms: Accuracy GoPubMed does NOT report MeSH terms as assigned by NLM!!! Major di...
Example GoPubMed Record: Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011
Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011
Curating  GoPubMed Record: Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011
Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011
‘ Won’ by default   Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011 <ul><li>Variances f...
Conclusions  Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011 <ul><li>GoPubMed  &  PubRe...
Questions?  Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011 <ul><li>Sue Bradley, Health...
Upcoming SlideShare
Loading in...5
×

GoPubMed versus PubReminer comparison

2,217

Published on

Published in: Health & Medicine, Technology
0 Comments
1 Like
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total Views
2,217
On Slideshare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
26
Comments
0
Likes
1
Embeds 0
No embeds

No notes for slide

Transcript of "GoPubMed versus PubReminer comparison"

  1. 1. GoPubMed versus PubReMiner for analyzing PubMed search results: A head to head comparison of two free web ‘data mining’ tools Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011
  2. 2. Introduction Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011 <ul><li>Searching PubMed/MEDLINE requires MeSH term mapping & natural language analysis </li></ul><ul><li>Search strategies can be improved by examining textwords, MeSH terms etc., in relevant papers already found </li></ul><ul><li>GoPubMed & PubReMiner </li></ul><ul><ul><li>two free data-mining tools used to statistically analyze PubMed search results </li></ul></ul><ul><ul><li>statistical analyses of search fields (e.g. publication years, MeSH terms, author names) </li></ul></ul><ul><li>Some different fields covered (e.g. PubReMiner provides chemical substance name search, GoPubMed does not) </li></ul><ul><li>Head to head comparisons using PubMed citations & PMIDs on fields covered by both </li></ul>
  3. 3. <ul><li>PubReMiner is a data mining tool, mining PubMed - MEDLINE abstracts </li></ul><ul><li>Query entered as text, using PubMed format, or as PMIDs </li></ul><ul><li>Provides bibliometric statistical analysis of search results </li></ul><ul><li>Displayed in hyperlinked ‘frequency tables’ by year, journal, authors, textwords, MeSH terms, chemical substance names and country </li></ul>Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011
  4. 4. <ul><li>GoPubMed - a knowledge-based semantic browsing tool for life sciences </li></ul><ul><li>One of the first web 2.0 / semantic search engines </li></ul><ul><li>Uses two ontologies: GoGene & MeSH (& lists proteins) </li></ul><ul><li>Statistics feature is a “semanto-bibliometric analysis” of search results </li></ul>Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011
  5. 5. Methods Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011 <ul><li>A series of PubMed searches were conducted </li></ul><ul><li>Main search topic: “borderline personality disorder”[mh] AND “therapy”[mh] </li></ul><ul><li>Search developed to yield small number of records (n=129) </li></ul><ul><ul><li>Period covered: 2006-2010 </li></ul></ul><ul><li>PubMed IDs (PMIDs) entered into GoPubMed & PubReMiner </li></ul><ul><li>Statistical results were compared to those obtained by hand </li></ul><ul><li>Discrepancies in performance of tools was examined </li></ul>
  6. 6. Results <ul><li>Publication year </li></ul><ul><li>PubReMiner (PRM) displays chronologically </li></ul><ul><li>GoPubMed (GPM) displays by frequency </li></ul>PRM: GPM: No discrepancies found using either tool vs Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011
  7. 7. Author Names <ul><li>Giustini D*[au] entered in PubMed: 8 records found – PMIDs entered </li></ul><ul><li>PubReMiner found all </li></ul><ul><li>GoPubMed only found 7 – split into three entries </li></ul>PRM: GPM: vs Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011 ∑ =8 ∑ =7
  8. 8. Country <ul><li>PRM: </li></ul>PRM: vs GPM: Minor discrepancies found with both tools: a ‘tie’ Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011
  9. 9. Journal names <ul><li>GoPubMed enters different journal abbreviations for same journal </li></ul><ul><li>eISSN/ISSN given different short forms </li></ul><ul><li>Not all are correct NLM title abbreviations </li></ul>PRM: GPM:
  10. 10. MeSH terms <ul><li>Presented very differently in PubReminer & GoPubMed </li></ul><ul><li>PRM includes MeSH subheadings: </li></ul>
  11. 11. <ul><li>GoPubMed presents a neater table; no subheadings </li></ul><ul><li>But you may need to click through many, many pages </li></ul>MeSH terms Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011
  12. 12. PubReminer vs Actual PubMed Results MeSH terms: Accuracy Two indexed records were not included : PMIDs: 20579633 / 20107104 Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011
  13. 13. <ul><li>MeSH terms & chemical substance names NOT read for two PMIDs: 20579633 20107104 </li></ul>Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011
  14. 14. GoPubMed vs Actual PubMed Results MeSH terms: Accuracy GoPubMed does NOT report MeSH terms as assigned by NLM!!! Major discrepancies observed
  15. 15. Example GoPubMed Record: Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011
  16. 16. Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011
  17. 17. Curating GoPubMed Record: Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011
  18. 18. Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011
  19. 19. ‘ Won’ by default Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011 <ul><li>Variances found in results for every field covered by both tools (except publication year) </li></ul><ul><li>Problems found in results for both tools </li></ul><ul><li>PubReMiner found to produce better results for fields covered by both </li></ul>N/A = not available with this tool Summary of GoPubMed and PubReMiner comparisons
  20. 20. Conclusions Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011 <ul><li>GoPubMed & PubReMiner are useful supplements for devising searches </li></ul><ul><li>Use to determine common terms, keywords & MeSH, top authors, journals </li></ul><ul><li>Use GoPubMed & PubReMiner with caution in developing search strategies </li></ul><ul><ul><li>Examine bibliographic records, indexing, ‘related articles’ features in PubMed </li></ul></ul><ul><li>Watch developments in PubReMiner & GoPubMed </li></ul>
  21. 21. Questions? Sue Bradley, Dean Giustini – CHLA/ABSC Annual Conference Calgary, Alberta May 2011 <ul><li>Sue Bradley, Health Librarian, Consultant, Vancouver BC </li></ul><ul><li>Dean Giustini, UBC Biomedical Branch Library </li></ul>Creative Commons Attribution 2.5 Canada Licence
  1. A particular slide catching your eye?

    Clipping is a handy way to collect important slides you want to go back to later.

×