Skip to contents

Get Subject-Specific Parameters from an NLME Demand Model

Source: R/mixed-methods.R
get_subject_pars.beezdemand_nlme.Rd

Subject-level demand parameters for a beezdemand_nlme fit, matching the column / scale / expanded contract of get_subject_pars.beezdemand_tmb. Combines the population fixed effects with each subject's random-effect deviations and back-transforms to the natural scale.

Usage

# S3 method for class 'beezdemand_nlme'
get_subject_pars(object, expanded = NULL, ...)

Arguments

object

A beezdemand_nlme object.

expanded

Controls the return shape for fits with within-id-varying design columns (within-subject factors, within-id covariates, or multi-block pdBlocked specs).

  • NULL (default): auto-detect. Expands to one row per (subject, factor-level) cell when within-id variation is present; otherwise returns the wide one-row-per-subject shape.

  • TRUE: always attempt expansion (no-op when there is no within-id variation).

  • FALSE: always return the wide shape; emits a one-line warning when within-id variation is present (the affected subjects' Q0, alpha, Pmax, Omax are NA).

...

Currently unused.

Value

A data frame. Wide form: id, b_i, c_i (if alpha has random effects), Q0, alpha, Pmax, Omax. Expanded form additionally includes the within-subject factor column(s) with one row per (subject, factor-level) cell. Q0, alpha, Pmax, Omax are on the natural scale.

Random-effect aliases (b_i / c_i)

b_i / c_i are the subject's first-block random-effect deviation for Q0 / alpha. For parity with the TMB method these are reported on the natural-log linear-predictor scale: for the default param_space = "log10" the stored log10 deviation is multiplied by log(10); for param_space = "natural" the deviation is returned on the natural parameter scale. The full per-coefficient random effects remain available via ranef().

See also

get_subject_pars.beezdemand_tmb