I'm training a number of models with the aim of identifying those which will perform well on my data. As such, I'm using a lot of models that I am unfamiliar with, and they all have their own tuning parameters.

In the past, I have established effective parameter values for my models using an exhaustive sweep. This method can be prohibitive for models I am not familiar with, if they have several tuning parameters processing time can become even more of a barrier.

As I lack the breadth of experience, I could use some advice as to how this kind of issue should be approached, I generally use the caret package of R for my modelling purposes, due to the common interface. However, I cannot find any details outside of modelLookup() and caret documentation, which does not provide very much model specific information.


To semi-answer my own question. In caret, most models are implemented through a specific package.

i.e xgbTree method in the caret package of R, originates from the xgboost package

by referring to the documentation of the model itself, much more information will be provided regarding the parameters, what they do and how to use it.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.