'Classifier trained with different number of folds in GridSearchCV gives the same decision_fuction?
As stated in the title, I’m confused by the k-folding approach in GridSearchCV which allows you to specify its cv attribute as the number of folds. However, I trained three classifiers (AdaBoost) with 3, 5 and 7 folds respectively. When I use them to produce predictions on the same sample, they give identical results.
I think there should at least be some statistical variation. Could anyone please explain how exactly GridSearchCV handles k-folding?
Sources
This article follows the attribution requirements of Stack Overflow and is licensed under CC BY-SA 3.0.
Source: Stack Overflow
| Solution | Source |
|---|