Skip to contents

Compute starting values for optimization

Source: R/simulate.R
compute_start.Rd

Obtains rough starting values for the beta regression parameters by fitting a quasi-binomial GLM on the midpoint response. This provides a reasonable initialization for the interval likelihood optimizer.

Usage

compute_start(
  formula,
  data,
  link = "logit",
  link_phi = "logit",
  ncuts = 100L,
  lim = 0.5,
  repar = 2L
)

Arguments

formula

A Formula object (possibly multi-part).

data

Data frame.

Mean link function name.

Dispersion link function name.

ncuts

Number of scale categories.

lim

Uncertainty half-width.

Value

Named numeric vector of starting values.