search

microsoft / TREC-2019-Deep-Learning

Website for the TREC Deep Learning Track 2019
https://microsoft.github.io/TREC-2019-Deep-Learning/
Creative Commons Attribution 4.0 International
87 stars 28 forks source link

How did you calculate the score for document and query? #5

Closed Dinesh-Mali closed 5 years ago

Dinesh-Mali commented 5 years ago

In top 100 doc for train queries, you have given rank and related score. How did you calculated that score? is a score based on rank only, or is it based on document and query relevance? which features have you considered for estimating the scores and ranking for documents? Please guide, thanks.

eminey commented 5 years ago

Hi Dinesh,

We used a very basic IR model to generate the rankings and the scores.

Best, Emine

On Thu, Jul 4, 2019 at 10:56 AM Dinesh Mali notifications@github.com wrote:

In top 100 doc for train queries, you have given rank and related score. How did you calculated that score? is a score based on rank only, or is it based on document and query relevance? which features have you considered for estimating the scores and ranking for documents? Please guide, thanks.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/microsoft/TREC-2019-Deep-Learning/issues/5?email_source=notifications&email_token=AEOME3HAHXTPIEB43PXMYILP5XCM7A5CNFSM4H5XCAUKYY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4G5KYY3Q, or mute the thread https://github.com/notifications/unsubscribe-auth/AEOME3A4NJDA6KJKVDJPTT3P5XCM7ANCNFSM4H5XCAUA .

Dinesh-Mali commented 5 years ago

Thanks @eminey .