Check Collinearity Using VIF for Fitted Models — check

Computes Variance Inflation Factors (VIF) for fitted models returned by uni_reg(), multi_reg(), uni_reg_nbin(), or multi_reg_nbin(). Returns one VIF table per model. For multivariate models only

Usage

check_collinearity(model)

Arguments

model: A fitted model object with class "uni_reg", "multi_reg", "uni_reg_nbin", or "multi_reg_nbin".

Value

A tibble containing VIF values and interpretation. For multivariable models, returns one tibble. For univariate models, an error is raised indicating VIF is not applicable.

Examples

if (requireNamespace("gtregression", quietly = TRUE) &&
  requireNamespace("mlbench", quietly = TRUE) &&
  getRversion() >= "4.1.0") {
  data(PimaIndiansDiabetes2, package = "mlbench")
  pima <- PimaIndiansDiabetes2 |> dplyr::filter(!is.na(diabetes))
  pima$diabetes <- ifelse(pima$diabetes == "pos", 1, 0)
  fit <- multi_reg(pima,
    outcome = "diabetes",
    exposures = c("age", "mass", "glucose"),
    approach = "logit"
  )
  check_collinearity(fit)
}
#> # A tibble: 3 × 3
#>   Variable   VIF Interpretation 
#>   <chr>    <dbl> <chr>          
#> 1 age       1.04 No collinearity
#> 2 mass      1.02 No collinearity
#> 3 glucose   1.02 No collinearity