Zach Deane-Mayer
Zach Deane-Mayer
Darn. I thought I'd fixed that. Ok, multi class for caretStack should be easier implement than for caretEnsemble. I'll take a look. No promises on when I can get to...
Lets say x is your other model (the one you're ensembling xgboost with): ``` R new_train_control
[As explained in the vignette: "you should note that caretList does not support custom caret models"](https://htmlpreview.github.io/?https://github.com/zachmayer/caretEnsemble/blob/master/inst/doc/caretEnsemble-intro.html) You have to fit the second model with a second call to train. E.g.:...
Good! Glad to help
crap. that's a bug. If you can make a simple test case, I'll add it to the queue to fix.
@tsuresh83 Could you make a PR on github with the new version of your function so I can see what you changed? If the fix looks good to me and...
Nope, but I welcome pull requests! At the moment we're in the middle of some code cleanup and bug fixes, which will probably be followed by a CRAN release. Then...
`Observed values for each component model are not the same. Please re-train the models with the same Y variable` caretEnsemble currently requires all the models have the same target variable....
Here is one idea, based on a discussion on stackoverflow: http://stackoverflow.com/questions/28198103/alternative-to-do-call-for-large-datasets/28199260#28199260