r-caret: custom model with n-dimensional output

问题:

I am trying to fit an autoencoder model using caret. I run into problems because my output will be more than 1-dimensional (e.g. y in the call to train() will be a matrix in my case). Is there a way to account for output of higher dimension in caret?

回答1:

Right now, it only handles univariate outcomes. I'm not sure why you would be using train for an autoencoder though; those are usually used as a pre-processing step rather than the outcome itself.