-
Notifications
You must be signed in to change notification settings - Fork 14
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
BUG: Off-by-1 error in reporting stopping / iteration counting? #32
Comments
After playing around, I think the text error should just say "after doing". Will PR. |
nickeubank
added a commit
to nickeubank/DAME-FLAME-Python-Package
that referenced
this issue
Feb 23, 2021
Closed
Merged
nehargupta
added a commit
that referenced
this issue
Feb 21, 2023
* remove db code * fixed off by one errors Iterations start at 0. An iteration is completed once we have successfully attempted the match for that iteration. Also fixed pe_stopping and adjusted defaults for pre_dame and early_stop_iterations to be float('inf') * eliminated manual call to unittest * removed flame_db tests * update documentation early_stop_pe_frac is now eliminated and documentation is updated as it pertains to iterations. --------- Co-authored-by: Vittorio Orlandi <52432620+vittorioorlandi@users.noreply.github.com> Per @vittorioorlandi 1. Fixed the definition of early stopping with respect to PE frac. Previously it was comparing to the previous iteration PE when it should be comparing to the PE at baseline. Accordingly, the PE at the 0'th iteration is always the PE using all covariates — even if we make no exact matches. 2. Have everything starting at 0, with a completed iteration corresponding to successfully attempting matching. Updated defaults for pre_dame and early_stop_iterations; they are now float(‘inf’). This makes sense if they are supposed to be the number of iterations before something happens; if they are infinite they never happen (ie never early stop because of iteration and never switch to dame). 3. Updated some notebooks to reflect the above. 4. Eliminated the database folder and corresponding test folder for it -- this is temporary to get it up on PyPi with all the recent updates but will be put back shortly. 5. Updates documentation to reflect the above This takes care of #47 and #48 without breaking #32 or #33
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
I think the iteration counter may be using human indices (starting at 1) while the arguments use 0-counting, leading to confusing printouts (compare top and last lines?):
The text was updated successfully, but these errors were encountered: