You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We can get the y_cv_predicts as follows, but in this case there is no advantage to using SearchCV.
cv_predicts = []
for each_params in pg:
estimator_ = clone(gscv.estimator)
estimator_.set_params(**each_params)
cv_pred = cross_val_predict(estimator_, X, y, cv=gscv.cv)
cv_predicts.append(cv_pred)
score = scorer(y, cv_pred)
cv_predicts = np.array(cv_predicts).T
The benefit of adding this method is that it makes it easier to get the results of all configures.
It don't know it's acceptable or not to add a new method just for one purpose, but it is worth discussing because the side effects on other methods is likely to be small.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
We can get the prediction result of the best_estimator_ after using GridSearchCV.
My suggestion is to add a way to get all estimator_ prediction results when searching with GridSearchCV.
One of the reasons for adding it is to use it for the bias correction proposed in the following papers. Bootstrapping the Out-of-sample Predictions for Efficient and Accurate Cross-Validation
A bias correction for the minimum error rate in cross-validation
( The second paper can be implemented using gscv.cv_results_.)
We can get the y_cv_predicts as follows, but in this case there is no advantage to using SearchCV.
The benefit of adding this method is that it makes it easier to get the results of all configures.
It don't know it's acceptable or not to add a new method just for one purpose, but it is worth discussing because the side effects on other methods is likely to be small.
Beta Was this translation helpful? Give feedback.
All reactions