Skip to contents

Posterior fitted values for an stLMM fit

fitted.stLMM.Rd

Computes posterior fitted values for the original data used to fit an stLMM model. These fitted values include all model components that were fit: formula offsets, fixed effects, explicit grouped random effects, and saved or recovered structured latent process effects. Binomial fitted values default to posterior fitted probabilities, and negative-binomial fitted values default to posterior fitted mean counts.

Usage

# S3 method for class 'stLMM'
fitted(object, summary = TRUE,
  sub_sample = list(start = 1L, thin = 1L), scale = c("response", "link"),
  ...)

Arguments

object

An object returned by stLMM or recover.

summary

Logical; if TRUE, return posterior mean fitted values. If FALSE, return the posterior fitted-value sample matrix.

sub_sample

List with optional start and thin entries used to subset posterior draws. For models with structured process terms, this subsets the saved or recovered latent-process draw indices in object$recover_iter.

scale

For family = "binomial", return fitted probabilities on the "response" scale or the linear predictor on the "link" scale. For family = "negative_binomial", return fitted mean counts on the "response" scale or the log mean on the "link" scale. For Gaussian models these scales are identical.

...

Currently unused.

Value

If summary = TRUE, a numeric vector of posterior mean fitted values for the original fitted data rows, including rows whose response was missing at fit time. If summary = FALSE, a matrix with posterior draws in rows and original data rows in columns. The matrix has a draw_index attribute giving the original MCMC iteration indices used and, for summary = FALSE Polya-Gamma fits, a scale attribute.

Details

This method does not predict at new locations. For models with structured process terms, fitted values require saved or recovered latent process samples. For Polya-Gamma process models, stLMM() saves in-chain process draws by default as save_process = list(start = 1, thin = 1) and recover() subsets them. For Gaussian process models, call recover() on the fitted object to make those samples available.

Examples

set.seed(1)
dat <- data.frame(
  y = rnorm(8),
  x = rnorm(8),
  time = seq_len(8)
)

fit <- stLMM(
  y ~ x + ar1(time),
  data = dat,
  n_samples = 8,
  priors = list(
    resid = list(tau_sq = ig(2, 1)),
    ar1_1 = list(sigma_sq = ig(2, 1), phi = uniform(-0.8, 0.8))
  ),
  warmup = FALSE,
  verbose = FALSE
)
rec <- recover(fit)
mu_hat <- fitted(rec)
mu_draws <- fitted(rec, summary = FALSE)