Extract fitted missing-predictor summaries

imputed() reports the fitted values used for explicitly modelled missing predictors. Gaussian missing predictor values are reported as conditional modes from the fitted TMB likelihood. When TMB::sdreport() is available, std_error contains the corresponding likelihood-based conditional standard error for Gaussian missing predictor values. Binary missing predictor values are reported as fitted conditional probabilities from the Bernoulli/logit predictor model and the Gaussian response likelihood. Ordered categorical missing predictor values are reported as fitted conditional expected scores from the cumulative-logit predictor model and the Gaussian response likelihood. Unordered categorical missing predictor values are reported as fitted conditional modal category scores from the baseline-category softmax predictor model and the Gaussian response likelihood. Beta/proportion, zero-one beta boundary-proportion, and denominator-aware beta-binomial missing predictor values are reported as fitted conditional means from the fitted predictor model and the Gaussian response likelihood. Count missing predictor values are reported as fitted conditional expected counts from Poisson, negative-binomial, or zero-truncated negative-binomial predictor models and Gaussian response likelihood. Lognormal, Gamma, and Tweedie missing predictor values are reported as fitted conditional quadrature means from the positive or semi-continuous predictor model and Gaussian response likelihood. The first finite-state, beta/proportion, boundary-proportion, beta-binomial, count, lognormal, Gamma, and Tweedie routes report NA standard errors.

Usage

imputed(object, ...)

# S3 method for class 'drmTMB'
imputed(object, variable = NULL, rows = c("missing", "all"), se = TRUE, ...)

Arguments

object: A drmTMB fit.
...: Reserved for future extractor options.
variable: Optional missing-predictor name. The default uses the only modelled missing predictor in the fit.
rows: Which rows to return. "missing" returns only fitted missing predictor values. "all" returns retained model rows, with observed predictor values labelled as observed.
se: Logical; include conditional standard errors when the fit contains a successful TMB::sdreport() result.

Value

A data frame with variable, original_row, model_row, observed, estimate, std_error, source, and uncertainty_status.

Details

This is not multiple imputation: the output does not contain posterior means, posterior intervals, credible intervals, or pooled-imputation summaries.

Examples

set.seed(20260532)
n <- 48
dat <- data.frame(
  moisture = seq(-1.5, 1.5, length.out = n),
  canopy = cos(seq_len(n) / 5)
)
dat$body_mass_full <- 0.2 + 0.7 * dat$moisture - 0.2 * dat$canopy +
  rnorm(n, sd = 0.08)
dat$growth <- 0.6 + 1.1 * dat$body_mass_full - 0.3 * dat$moisture +
  rnorm(n, sd = 0.20)
dat$body_mass <- dat$body_mass_full
dat$body_mass[c(7, 19, 34, 43)] <- NA_real_

fit <- drmTMB(
  bf(growth ~ moisture + mi(body_mass), sigma ~ 1),
  family = gaussian(),
  data = dat,
  impute = list(body_mass = body_mass ~ moisture + canopy),
  missing = miss_control(predictor = "model")
)
imputed(fit)
#>    variable original_row model_row observed   estimate  std_error
#> 1 body_mass            7         7    FALSE -0.5977839 0.06694745
#> 2 body_mass           19        19    FALSE  0.1230898 0.06692573
#> 3 body_mass           34        34    FALSE  0.4920569 0.06762264
#> 4 body_mass           43        43    FALSE  1.1676530 0.06729264
#>             source uncertainty_status
#> 1 conditional_mode                 ok
#> 2 conditional_mode                 ok
#> 3 conditional_mode                 ok
#> 4 conditional_mode                 ok