Skip to contents

Print, summarize, and plot stLMM objects

stLMM-methods.Rd

S3 methods for compact inspection of fitted, recovered, and prediction objects. Summaries report posterior means, standard deviations, and quantiles for active sample blocks only. Plot methods use base graphics and choose conservative defaults to avoid drawing very large posterior or prediction objects by accident.

Usage

# S3 method for class 'stLMM'
print(x, ...)
# S3 method for class 'stLMM'
summary(object, probs = c(0.025, 0.5, 0.975),
  burn = 0L, parameters = NULL, ...)
# S3 method for class 'stLMM'
plot(x, type = c("trace", "density"), parameters = NULL,
  max_parameters = 12L, burnin = 0L, thin = 1L, ...)
# S3 method for class 'stLMM_chains'
print(x, ...)
# S3 method for class 'stLMM_chains'
summary(object, probs = c(0.025, 0.5, 0.975),
  diagnostics = TRUE, include_w = FALSE, burn = 0L,
  parameters = NULL, ...)
# S3 method for class 'stLMM_chains'
plot(x, type = c("trace", "density"),
  parameters = NULL, max_parameters = 12L, include_w = FALSE,
  n_col = NULL, chain_colors = NULL, lwd = 1, burnin = 0L,
  thin = 1L, ...)
# S3 method for class 'stLMM_recovery'
print(x, ...)
# S3 method for class 'stLMM_recovery'
summary(object, probs = c(0.025, 0.5, 0.975),
  burn = 0L, parameters = NULL, include_w = FALSE, max_w = 20L, ...)
# S3 method for class 'stLMM_recovery_chains'
summary(object,
  probs = c(0.025, 0.5, 0.975), diagnostics = TRUE,
  include_w = FALSE, burn = 0L, parameters = NULL, ...)
# S3 method for class 'stLMM_recovery'
plot(x, term = NULL, nodes = NULL,
  type = c("trace", "density", "fitted"), observed = NULL,
  max_nodes = 12L, burnin = 0L, thin = 1L, ...)
# S3 method for class 'stLMM_recovery_chains'
plot(x, type = c("trace", "density"),
  parameters = NULL, max_parameters = 12L, include_w = FALSE, ...)
# S3 method for class 'stLMM_prediction'
print(x, ...)
# S3 method for class 'stLMM_prediction'
summary(object, probs = c(0.025, 0.5, 0.975),
  include_y = !is.null(object$y_samples), ...)
# S3 method for class 'stLMM_prediction'
plot(x, sample = c("mu", "y"),
  type = c("interval", "density", "scatter"), rows = NULL,
  observed = NULL, probs = c(0.025, 0.5, 0.975), max_rows = 200L,
  burnin = 0L, thin = 1L, ...)
# S3 method for class 'stLMM_prediction_chains'
plot(x, sample = c("mu", "y"),
  type = c("interval", "density"), rows = NULL,
  probs = c(0.025, 0.5, 0.975), max_rows = 200L,
  burnin = 0L, thin = 1L, ...)

Arguments

x, object

An stLMM, stLMM_recovery, or stLMM_prediction object, or the corresponding multi-chain object.

probs

Posterior quantiles to report. Prediction interval plots expect three probabilities: lower, center, and upper.

type

Plot type. For fit objects, "trace" or "density". For recovery objects, "trace", "density", or "fitted". For prediction objects, "interval", "density", or "scatter".

parameters

Optional exact parameter names to summarize or plot for an stLMM fit. For summaries, unlisted parameters are omitted from the returned parameter table or tables.

max_parameters

Maximum number of parameters to plot by default.

burn

Number of initial posterior draws to discard before computing fit summaries and multi-chain diagnostics. The default 0L uses all retained draws. This affects only the summary call; it does not modify the stored posterior samples.

diagnostics

Logical; for multi-chain summaries, compute coda diagnostics such as potential scale reduction and effective sample size.

burnin

Number of initial plotted draws to discard. This affects only the plotting call; it does not modify the stored posterior samples.

thin

Keep every thin-th plotted draw after burnin. This affects only the plotting call.

include_w

Logical; include summaries of saved or recovered latent process samples. Defaults to FALSE because these can be large.

n_col

Optional number of columns for multi-chain fit trace or density plot panels. Defaults to a compact base graphics layout.

chain_colors

Optional vector of colors for multi-chain fit plots. Values are recycled to the number of chains.

lwd

Line width for multi-chain fit trace or density plots.

max_w

Maximum number of latent process nodes summarized when include_w = TRUE.

term

Recovered process term to plot. Defaults to the first recovered term.

nodes

Latent process nodes to plot.

max_nodes

Maximum number of recovered nodes to plot by default.

include_y

Logical; include posterior predictive observation summaries when y_samples were simulated.

sample

Prediction sample matrix to plot: "mu" or "y".

rows

Prediction rows to plot.

observed

Observed values for recovery fitted-value scatter plots or prediction scatter plots. Recovery plots use the fitted response by default.

max_rows

Maximum number of prediction rows plotted by default.

...

Additional arguments passed to base plotting functions.

Value

Print methods return x invisibly. Summary methods return summary objects. Plot methods return x invisibly.