Skip to contents

Plot Diagnostics for MS-VARMA-GARCH Models

plot.ms_diagnostics.Rd

Creates diagnostic plots for EM algorithm convergence and parameter evolution.

Usage

# S3 method for class 'ms_diagnostics'
plot(
  x,
  type = c("all", "ll_evolution", "parameters", "sigma"),
  parameters = NULL,
  normalize = FALSE,
  quiet = FALSE,
  ...
)

Arguments

x

An object of class ms_diagnostics returned by the fitting procedure.

type

Character string specifying which plots to produce. One of:

"all"

Produce all diagnostic plots (default).

"ll_evolution"

Plot log-likelihood evolution across EM iterations.

"parameters"

Plot parameter evolution across EM iterations.

"sigma"

Plot conditional volatility (sigma) evolution.

parameters

Optional character vector of parameter names to include in the parameter evolution plot. If NULL (default), all parameters are plotted. Supports regex patterns when a single string containing regex metacharacters (^, $, ., *, [) is provided.

normalize

Logical. If TRUE, normalize parameter values to [0, 1] within each parameter for easier comparison across different scales. Default is FALSE.

quiet

Logical. If TRUE, suppress warnings from numeric coercion during parameter extraction. Default is FALSE.

...

Additional arguments (currently unused).

Value

Invisibly returns the ggplot object(s). Called primarily for side effects (printing plots).

Details

The function produces up to three types of diagnostic visualizations:

Log-Likelihood Evolution (type = "ll_evolution"): Two plots showing (1) the log-likelihood value after each M-step, and (2) the change in log-likelihood per iteration. Useful for assessing convergence and detecting any non-monotonic behavior.

Parameter Evolution (type = "parameters"): Faceted plot showing how each parameter evolves across EM iterations, with separate colors for each regime state. The parameters argument can filter to specific parameters of interest.

Sigma Evolution (type = "sigma"): Faceted plot showing the mean conditional volatility (with ± 1 SD ribbon) for each series across iterations, colored by state.

See also

summary.ms_diagnostics for text summaries of diagnostics.

Examples

if (FALSE) { # \dontrun{
# After fitting a model with diagnostics enabled
fit <- fit_ms_varma_garch(data, n_states = 2, collect_diagnostics = TRUE)
diagnostics <- attr(fit, "diagnostics")

# Plot all diagnostics
plot(diagnostics)

# Plot only log-likelihood evolution
plot(diagnostics, type = "ll_evolution")

# Plot specific parameters
plot(diagnostics, type = "parameters", parameters = c("alpha_1", "beta_1"))

# Plot parameters matching a regex pattern
plot(diagnostics, type = "parameters", parameters = "^alpha")

# Normalized parameter plot for cross-parameter comparison
plot(diagnostics, type = "parameters", normalize = TRUE)
} # }