In the function imputeWithAndWithoutEstimates, when estimating model including manual estimates a simpler set of model is used.
The included models are:
defaultMean
defaultLm
defaultExp
defaultNaive
defaultMixedModel
However, the similarity between these models are extremely high, and they are all global model except for the defaultNaive.
This defeats the purpose of the ensemble model, and defying one of the underlying requirement of the ensemble model where the set of model should be as diverse as possible.
Further, there is no reason to use a smaller set, thus we should revert to use the complete model.
In the function
imputeWithAndWithoutEstimates
, when estimating model including manual estimates a simpler set of model is used.The included models are:
However, the similarity between these models are extremely high, and they are all global model except for the
defaultNaive
.This defeats the purpose of the ensemble model, and defying one of the underlying requirement of the ensemble model where the set of model should be as diverse as possible.
Further, there is no reason to use a smaller set, thus we should revert to use the complete model.