CVfit: Cross-Validation for Generalized Estimating Equations (GEE)

Description

This function performs k-fold cross-validation for model selection in the context of Generalized Estimating Equations (GEE). It is designed to evaluate the performance of different models specified by a range of lambda values, choosing the one that minimizes the cross-validation criterion.

Usage

CVfit(
  formula,
  id,
  data,
  family,
  scale.fix,
  scale.value,
  fold,
  lambda.vec,
  pindex,
  eps,
  maxiter,
  tol,
  corstr = "independence",
  ncore = 1
)

Value

An object of class "CVfit", which is a list containing:

fold: The number of folds used in the cross-validation.
lam.vect: The vector of lambda values tested.
cv.vect: The cross-validation error for each lambda.
lam.opt: The lambda value that resulted in the minimum cross-validation error.
cv.min: The minimum cross-validation error.
call: The matched call.

Arguments

formula: an object of class "formula" (or one that can be coerced to that class): a symbolic description of the model to be fitted.
id: a vector which identifies the cluster/group for each observation.
data: an optional data frame containing the variables in the model.
family: a description of the error distribution and link function to be used in the model.
scale.fix: logical; if TRUE, the scale parameter is fixed to scale.value.
scale.value: the value of the scale parameter when scale.fix is TRUE.
fold: the number of folds to be used in the cross-validation.
lambda.vec: a vector of lambda values for which the cross-validation error will be calculated.
pindex: an optional numeric vector specifying a parameter index.
eps: the threshold for convergence criteria.
maxiter: the maximum number of iterations for the convergence of the algorithm.
tol: the tolerance level for the convergence of the algorithm.
corstr: the correlation structure used.
ncore: if greater than 1, the function will use parallel computation.

Details

Note that this is a re-implemented version with parallel computing.