classValidation: Validation of binary classification rule

In lampk/R306:

Description

Validation of binary classification rule

Usage

classValidation(model, data, type, scope = NULL,
  validation_method = "bootstrap", B = NULL, B_mat = NULL,
  test_set_method = ifelse(validation_method == "bootstrap", "all",
  "excluded"), show_selected_coef = F, criteria = c(Brier, c_index, calib),
  roc = F, fpr = seq(0, 1, length = 101), ...)

Arguments

model	Model's formula
data	Data to fit and assess model performance
type	Type of statistical model
scope	Scope in stepwise variable selection
validation_method	"cv" (cross-validation) or "bootstrap"
B	number of bootstrap samples / B-fold cross-validation
B_mat	Data matrix for cross-validation
test_set_method	How is the test set defined (either "all" -> use original dataset as test data (default for bootstrap), or "excluded" -> use patients excluded from cv/bootstrap sample as test data)
show_selected_coef	If TRUE, show the selected coefficients (only for glm's with model selection)
roc	If TRUE, true positive rates corresponding to false positive rates given in argument fpr are also evaluated and cv/bootstrap corrected
...	Other parameters for fit_method and fit_response

Value

List of validation output

lampk/R306 documentation built on May 20, 2019, 7:34 p.m.