Stepwise Beta regression by AICc (finite-sample corrected AIC) — betareg_step

Greedy forward/backward search minimizing AICc computed on betareg fits with optional precision-submodel selection and observation weights.

Usage

betareg_step_aicc(
  X,
  Y,
  direction = "both",
  link = "logit",
  link.phi = "log",
  type = "ML",
  trace = FALSE,
  max_steps = NULL,
  epsilon = 1e-08,
  X_phi = NULL,
  direction_phi = c("none", "both", "forward", "backward"),
  weights = NULL
)

Arguments

X: Numeric matrix (n × p) of mean-submodel predictors.
Y: Numeric response in (0,1). Values are squeezed to (0,1) internally.
direction: Stepwise direction for the mean submodel: "both", "forward", or "backward".
link: Link for the mean submodel (passed to betareg). Default "logit".
link.phi: Link for precision parameter. Default "log".
type: Likelihood type for betareg, e.g. "ML".
trace: Logical; print stepwise trace.
max_steps: Maximum number of greedy steps (default p).
epsilon: Minimal AICc improvement to accept a move.
X_phi: Optional matrix of candidate predictors for the precision (phi) submodel. When direction_phi enables precision updates and X_phi is NULL, the function reuses X.
direction_phi: Stepwise direction for the precision submodel. Defaults to "none" (no phi selection). Supported values mirror direction.
weights: Optional non-negative observation weights passed to betareg().

Value

See betareg_step_aic().

Examples

set.seed(1);
X <- matrix(rnorm(400), 100, 4);
Y <- plogis(X[,1]+0.5*X[,2])
betareg_step_aicc(X, Y)
#>     (Intercept)              X1              X2              X3              X4 
#>    3.248654e-04    9.858079e-01    4.932358e-01    0.000000e+00    0.000000e+00 
#> phi|(Intercept) 
#>    1.307252e+01 
Y <- rbeta(100, Y*25, (1-Y)*25);
betareg_step_aicc(X, Y)
#>     (Intercept)              X1              X2              X3              X4 
#>    -0.009987406     1.044515121     0.532545592     0.000000000     0.000000000 
#> phi|(Intercept) 
#>     3.135373326