Map a Qualitative Confidence Level to a Variance Inflation Factor (VIF)

Maps intuitive confidence levels to VIF values for easy prior specification.

Usage

confidence_to_vif(confidence = c("low", "medium", "high"))

Arguments

confidence

Character; one of "low", "medium", or "high".

Value

Numeric; VIF value (5.0 for low, 2.5 for medium, 1.5 for high).

Details

The mapping is:

low

VIF = 5.0; high uncertainty about \(K_J\)

medium

VIF = 2.5; moderate uncertainty

high

VIF = 1.5; high confidence (near Poisson boundary)

Higher confidence implies lower variance, which corresponds to lower VIF. The "high" setting (VIF = 1.5) is close to the A1 feasibility boundary.

See also

vif_to_variance for converting VIF to variance

Examples

# Get VIF for medium confidence
vif <- confidence_to_vif("medium")  # Returns 2.5

# Complete workflow
mu_K <- 5
vif <- confidence_to_vif("low")
var_K <- vif_to_variance(mu_K, vif)
fit <- DPprior_a1(J = 50, mu_K = mu_K, var_K = var_K)