cvCoRF: n-fold cross-validation of CoRF
In DennisBeest/CoRF: Co-data guided RandomForest (CoRF)

cvCoRF

R Documentation

n-fold cross-validation of CoRF

Conducts a n-fold cross-validation of a CoRF object to obtain an unbiased estimate of the predictive error. Primarly for when gamma is tuned.

cvCoRF(CoRFObject, nfold = 10, SeedCVO = 123, GammaSeq = NULL,
  ntree = NULL, SelectForest = TRUE)

`CoRFObject`	A CoRF object.
`nfold`	The cross-validation fold.
`SeedCVO`	The seed used to draw the cross-validation folds.
`GammaSeq`	Set gamma sequence to be used in the cross-validation. By default uses the same sequence as used to fit the CoRF object.
`ntree`	Set the number of trees used per RF in the cross-validation. By default uses the same number of trees as used to fit the CoRF object.

Returns An object with the result of the cross-validation, containng the following components.

`foldid`	The cross-validation folds used
`OptGammaPerFold`	When more than one gamma was used, this returns per fold the gamma with lowest brier/auc.
`Predictions`	The cross-validated predictions, for the base RF, the CoRF with gamma=1, and optionally the predictions for the CoRF selected by brier or auc.
`cvResult`	The result of the cross-validation, e.g. cross-validated auc, brier, and accuracy.