The correct answer is: A. stacking has less stable cv compared to blending.
Stacking is a meta-learning technique that combines the predictions of multiple base models to produce a more accurate prediction. The base models are trained on the same data, but they can be different models with different hyperparameters. The predictions of the base models are then combined using a weighted average, where the weights are determined by a training set that is held out from the training of the base models.
Blending is a similar technique, but the predictions of the base models are combined using a simple average, without any weights. This makes blending simpler than stacking, but it also means that it is less likely to produce an accurate prediction.
The reason why stacking has less stable CV compared to blending is that the weights of the base models in stacking are determined by the training set that is held out from the training of the base models. This means that the weights can be sensitive to small changes in the data. In contrast, the predictions of the base models in blending are simply averaged, so they are not as sensitive to small changes in the data.
Here is a more detailed explanation of each option:
- Option A: Stacking has less stable CV compared to blending. This is because the weights of the base models in stacking are determined by the training set that is held out from the training of the base models. This means that the weights can be sensitive to small changes in the data. In contrast, the predictions of the base models in blending are simply averaged, so they are not as sensitive to small changes in the data.
- Option B: In blending, you create out of fold prediction. This is not true. In both stacking and blending, you can use out of fold prediction to estimate the performance of the combined model.
- Option C: Stacking is simpler than blending. This is true. Stacking is a more complex technique than blending, but it is also more likely to produce an accurate prediction.
- Option D: None of these. This is not true. Option A is the correct answer.