cv_gpls: Cross Validation for Generalized Projection to Latent...

In abnormally-distributed/cvreg: Cross Validation and Robust Estimation Utilities

Description

Cross Validation for Generalized Projection to Latent Structures Regression

Usage

cv_gpls(
  formula,
  data,
  family = NULL,
  link = NULL,
  cv.method = "boot632",
  nfolds = 5,
  nrep = 4,
  crit = NULL,
  select = "oneSE",
  folds = NULL,
  preproc = c("center", "scale")
)

Arguments

formula	a model formula
data	a training data set
family	the glm family. One of "gaussian", "student", "poisson", "negbinom", "binomial", "multinom", "gamma", "invgauss"
link	the link function. See details for more information.
cv.method	preferably one of "boot632" (the default), "cv", or "repeatedcv".
nfolds	the number of bootstrap or cross-validation folds to use. defaults to 5.
nrep	the number of repetitions for cv.method = "repeatedcv". defaults to 4.
crit	the criterion by which to evaluate the model performance. See details for more information.
select	the selection rule to use. Should be one of "best" or "oneSE" (the default).
folds	a vector of pre-set cross-validation or bootstrap folds from caret::createResample or caret::createFolds.

Details

The available summary statistics for the argument "crit" depend on which likelihood function is chosen for the glm family.

If the outcome is multinomial, it should be one of "Kappa" (the default), "Accuracy", "Mean_F1", "Mean_Sensitivity", "Mean_Specificity", "Mean_Pos_Pred_Value", "Mean_Neg_Pred_Value", "Mean_Precision", "Mean_Recall", or "Mean_Detection_Rate".

If the outcome is binomial, it should be one of "Spec" (Specificity, the default) or "Sens".

Otherwise, it should be one of "MAE" (the default) or "MSE".


#' The following link functions are available for each distribution:

Gaussian & Student's T: "identity"
Binomial & Multinomial: "logit", "probit", "cauchit", "robit" (Student T with 3 df), and "cloglog"
Poisson & Negative Binomial: "log"
Gamma: "inverse" (1 / x)
Inverse Gaussian: "invsquare" (1/x^2)

Value

a train object

abnormally-distributed/cvreg documentation built on May 3, 2020, 3:45 p.m.