search

kosukeimai / fastLink

R package fastLink: Fast Probabilistic Record Linkage
272 stars 48 forks source link

aggconfusion development update #69

Closed SamShin closed 1 year ago

SamShin commented 1 year ago

Hi,

Thank you for writing such an amazing package. I am trying to calculate the precision and recall and so I am trying to use the confusion table. But since I want to use blocking methods, it seems like I would have to use aggconfusion to do so which is currently under development. Is there any other ways to find the precision and recall outside of manually calculating them?

Thanks in advance, Sam

aalexandersson commented 1 year ago

Disclaimer: I am a regular fastLink user, not a fastLink developer.

You can use the function confusion() to find the precision and recall. I provided an example in https://github.com/kosukeimai/fastLink/issues/63#issuecomment-1255559880.

Precision = Positive Predictive Value = True Positives / Links = TP / (TP + FP). Recall = Sensitivity = True Positives / Matches = TP / (TP + FN).

SamShin commented 1 year ago

Thank you so much for the amazing example! This is exactly what I needed.