Considerations for the search summary table

ESHackathon / CiteSource

http://www.eshackathon.org/CiteSource/

GNU General Public License v3.0

16 stars 1 forks source link

Considerations for the search summary table #165

Closed rootsandberries closed 2 months ago

rootsandberries commented 1 year ago

I'm working with our search summary table for a project, and have some thoughts for improvement:

[ ] Change footnotes 4 and 5 to indicate that these are percentages. So Number of final included citations / Number of distinct records * 100, or Percent of final included citations in total number of distinct records.
[ ] Change Sensitivity/Recall column heading to Sensitivity (aka Recall)
[ ] I wonder if the F-score might be useful to add as a last column as a metric to evaluate and compare the usefulness of each source?

TNRiley commented 1 year ago

1 and 2 are easy edits that I can take care of. I haven't worked with F score much, but can look into it in the future. Can you explain how it would be used?

TNRiley commented 4 months ago

@rootsandberries there had been some changes made to tables a while back. Are these issues still relevant to the current tables?

rootsandberries commented 4 months ago

The tables look great! The only metric that I thought might be a useful addition in the table is 'f-score'. You can see it used here in this paper. It kind of provides a nice single number you can use to compare across sources to get a sense of the best source in terms of balance between sensitivity and precision. Not necessary, but perhaps somewhat easy to implement?

@rootsandberries there had been some changes made to tables a while back. Are these issues still relevant to the current tables?

TNRiley commented 2 months ago

closing for now and adding F-score as a new issue for future