Catch unsupported model types in `BinaryThresholdPredictor`.

Open ablaom opened this issue 3 years ago • 1 comments

To resolve #486.

Also adds a docstring to resolve https://github.com/alan-turing-institute/MLJ.jl/issues/973.

Oct 21 '22 04:10 ablaom

Codecov Report

Merging #487 (9202c4b) into dev (3f25d25) will increase coverage by 0.10%. The diff coverage is 100.00%.

@@            Coverage Diff             @@
##              dev     #487      +/-   ##
==========================================
+ Coverage   77.32%   77.43%   +0.10%     
==========================================
  Files          16       16              
  Lines        1116     1121       +5     
==========================================
+ Hits          863      868       +5     
  Misses        253      253

Impacted Files	Coverage Δ
src/builtins/ThresholdPredictors.jl	`92.74% <100.00%> (+0.30%)`	:arrow_up:

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

Oct 21 '22 05:10 codecov-commenter

@OkonSamuel Can you please review this?

Oct 25 '22 22:10 ablaom

@OkonSamuel Thanks for the review.

I have replaced the synthetic data with the Pima Indian diabetes data, and replaced f1score with balanced_accuracy. In this case one sees a "statistically significant" improvement in the objective function.

Oct 26 '22 23:10 ablaom