To make split conformal predictions, you have to pre-train a model on X_train, then use the fit method with the cv=prefit parameter with MapieRegressor on X_calib. This PR proposes to do it automatically when using the fit method with the cv=1 parameter on X = X_train U X_calib.
Fixes #3
Type of change
New feature (non-breaking change which adds functionality)
This change requires a documentation update
How Has This Been Tested?
[x] Modify tests in test_regression.py file to verify the changes when using cv=1 (in particular: coverage, width)
Description
To make split conformal predictions, you have to pre-train a model on
X_train
, then use thefit
method with thecv=prefit
parameter withMapieRegressor
onX_calib
. This PR proposes to do it automatically when using thefit
method with thecv=1
parameter onX = X_train U X_calib
.Fixes #3
Type of change
How Has This Been Tested?
test_regression.py
file to verify the changes when usingcv=1
(in particular: coverage, width)Checklist
make lint
make type-check
make tests
make coverage
make doc