quepid
quepid copied to clipboard
Tooltip text for Scorers display
Is your feature request related to a problem? Please describe. Scorers view could use documentation of the scorers to make it easier for users to choose which one to use.
Describe the solution you'd like Write a one sentence blurb for each scorer, to be displayed as a tooltip. Describe alternatives you've considered none Additional context none
I need to do some research to figure out how to have tooltips in Quepid's UI, this text spans the expected tooltip content.
The metrics are:
These three are gain based metrics. Each new relevant document contributes some value when it appears in the ranked list.
nDCG@10 -- Normalized discounted sum of the gain values. The ratio of the DCG to the Ideal DCG that results from the optimal ordering of the rated relevant documents. Useful for comparing across queries. Useful for averaging over multiple queries. DCG@10 -- Discounted sum of the gain values. Not useful for comparing across queries. Not useful for averaging over multiple queries. CG@10 -- Simple sum of the gain values. Not useful for comparing across queries. Not useful for averaging over multiple queries.
P@10 -- The effective accuracy, of the top k (10) retrieved, how many of them were in the set of documents rated relevant. Useful for comparing across queries. Useful for averaging across queries.
AP (while still called @10, AP is computed on |R| the set of all relevant documents) The average of the precision points (accuracy) at each rank at which a relevant document appears. This average is across the full set of relevant documents. Useful for comparing across queries. Useful for averaging over multiple queries. This metric is the most common summary metric for capturing overall performance of a retrieval system independent of the retrieval task.
RR (MRR when averaged across queries), not yet merged into Quepid. The reciprocal of the rank of the first relevant document to occur in the ranked list. This metric is appropriate for evaluating known item, or single target item search. Somewhat useful for comparing across queries. Useful for averaging across queries.
v1-520 -- average rating 100 discounted by edit distance from best @ 10 (why is this in quepid?) v1-527 -- is exactly the same code as v1-520 and should be removed as redundant.
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Please do advocate for this issue and ideally submit a patch to get the attention of the maintainers!
bad stale bot! We still want this! Sigh, better get this done someday ;-)