search

dongfang91 / Generate-and-Rank-ConNorm

Generate-and-Rank Framework with Semantic Type Regularization for Biomedical Concept Normalization
25 stars 4 forks source link

Evaluation metric #3

Open Mahmedturk opened 4 years ago

Mahmedturk commented 4 years ago

Hi,

The accuracy reported is the classification accuracy or it is the percentage of entities correctly normalized over total # of entities?

dongfang91 commented 4 years ago

It is the percentage of entities correctly normalized over the total # of entities.

Mahmedturk commented 4 years ago

Do you consider it correct if only the first candidate in the list is correct or you consider it correct if correct one falls in top 10 candidates?

dongfang91 commented 4 years ago

We consider it correct only if the gold concept is the first candidate in the list.