caret (version 5.13-037)

modelLookup: Descriptions Of Models Available in train()

Description

This function enumerates the parameters and characteristics for models used by train

Usage

modelLookup(model = NULL)

Arguments

model
a character string associated with the method argument of train. If no value is passed, all models are returned

Value

  • a data frame with columns
  • modela character string for the model code
  • parameterthe tuning parameter name
  • labela tuning parameter label (used in plots)
  • seqa logical; can sub-models be used to decrease training time (see the Details section)
  • forRega logical; can the model be used for regression?
  • forClassa logical; can the model be used for classification?
  • probModela logical; does the model produce class probabilities?

Details

One characteristic listed in the output is whether or not sub-models can be used for prediction. For example, if a PLS model is fit with X components, PLS models with train exploits this characteristic whenever possible.

These types of tuning parameters are defined as "sequential" parameters since one value can be used to derive a sequences of predictions. Examples of model codes that include sequential tuning parameters are; blackboost, ctree, earth, enet, foba, gamboost, gbm, glmboost, glmnet, lars, lars2, lasso, logitBoost, pam, partDSA, pcr, pls, relaxo, rpart, scrda and superpc.

See Also

train

Examples

Run this code
modelLookup()

modelLookup("gbm")

Run the code above in your browser using DataCamp Workspace