Get Demand Parameter Estimated Marginal Means for TMB Model

Computes estimated marginal means (EMMs) for demand parameters from a beezdemand_tmb model. Uses design matrices and beta vectors with vcov from TMB::sdreport().

Usage

# S3 method for class 'beezdemand_tmb'
get_demand_param_emms(
  fit_obj,
  param = c("Q0", "alpha"),
  factors_in_emm = NULL,
  at = NULL,
  ci_level = 0.95,
  ...
)

Arguments

fit_obj

A beezdemand_tmb object.

param

Character. Which parameter to compute EMMs for: "Q0" or "alpha".

factors_in_emm

Character vector of factors to retain in the EMM reference grid. If it names a strict subset of the fitted factors, the omitted factors are marginalized over using equal weights across the full crossing of their levels (emmeans' default weights = "equal"), matching the NLME backend. If NULL (default), all fitted factors are retained (no marginalization). Under asymmetric collapse_levels you may name either the original factor or its collapsed per-parameter column; a name that resolves to neither for this parameter is rejected with an error.

at

Named list specifying factor levels and continuous-covariate values for conditional EMMs. For continuous covariates, a single numeric value per covariate; multiple values produce a warning and only the first is used. at on a marginalized (omitted) factor restricts the level set averaged over.

ci_level

Numeric. Confidence level for intervals.

...

Additional arguments.

Value

A tibble with columns: level, estimate, std.error, conf.low, conf.high.

Note

Marginalization is exact because Q0/alpha are linear in the fixed-effect coefficients on the log scale, so averaging the reference-grid design rows and then multiplying by the coefficient vector equals averaging the per-cell parameter predictions.

Examples

# \donttest{
data(apt_full)
# 40 subjects per gender keep the example fast; use the full data in practice
ids <- unique(apt_full[c("id", "gender")])
ids <- ids[ids$gender %in% c("Male", "Female"), ]
keep <- unlist(lapply(split(ids$id, ids$gender), head, 40))
dat <- apt_full[apt_full$id %in% keep, ]
fit <- fit_demand_tmb(dat, equation = "exponential",
                      factors = "gender", verbose = 0)
#>   equation='exponential': Dropped 501 zero-consumption observations (859 remaining).
get_demand_param_emms(fit, param = "Q0")
#> # A tibble: 2 × 6
#>   level         estimate estimate_log std.error conf.low conf.high
#>   <chr>            <dbl>        <dbl>     <dbl>    <dbl>     <dbl>
#> 1 gender=Female     4.49         1.50     0.113     3.59      5.61
#> 2 gender=Male       7.74         2.05     0.110     6.24      9.60
get_demand_param_emms(fit, param = "alpha")
#> # A tibble: 2 × 6
#>   level         estimate estimate_log std.error conf.low conf.high
#>   <chr>            <dbl>        <dbl>     <dbl>    <dbl>     <dbl>
#> 1 gender=Female  0.0102         -4.59     0.169  0.00730    0.0142
#> 2 gender=Male    0.00937        -4.67     0.154  0.00693    0.0127
# }