Base class for RTMB Fit objects

Details

An R6 base class providing common methods for Bayesian and MAP inference results.

Public fields

model: The `RTMB_Model` object used for estimation.

Methods

Public methods

RTMB_Fit_Base$get_point_estimate()
RTMB_Fit_Base$estimate()
RTMB_Fit_Base$EAP()
RTMB_Fit_Base$MAP()
RTMB_Fit_Base$rotate()
RTMB_Fit_Base$posterior_predict()
RTMB_Fit_Base$pp_check()
RTMB_Fit_Base$fa_rotate()
RTMB_Fit_Base$clone()

Method `get_point_estimate()`

Abstract method to get a point estimate for a target parameter.

Usage

RTMB_Fit_Base$get_point_estimate(target, ...)

Arguments

target: Character string specifying the target parameter name.
...: Additional arguments.

Returns

A numeric array or matrix of the point estimate.

Method `estimate()`

Get point estimates for parameters, transformed parameters, and generated quantities.

Usage

RTMB_Fit_Base$estimate(
  pars = NULL,
  type = c("mean", "EAP", "marginal_map", "joint_map", "MAP"),
  component = c("all", "parameters", "transform", "generate"),
  chains = NULL,
  best_chains = NULL,
  drop = TRUE,
  ...
)

Arguments

pars: Optional character or numeric vector of parameter names or indices to extract. Supports special keywords: "parameters", "transform", "generate", and "all".
type: Character string specifying the estimation type.
component: Character string specifying the component to filter by.
chains: Numeric vector of chains to include.
best_chains: Number of best chains to include.
drop: Logical; if TRUE and only one parameter is selected, return the value directly instead of a list.
...: Additional arguments passed to draws().

Returns

A named list of point estimates, or a single value if `drop = TRUE`.

Method `EAP()`

Calculate Expected A Posteriori (EAP) estimates from posterior samples.

Usage

RTMB_Fit_Base$EAP(
  pars = "parameters",
  chains = NULL,
  best_chains = NULL,
  drop = TRUE,
  ...
)

Arguments

pars: Optional character vector of parameter names to extract.
chains: Numeric vector of chains to include.
best_chains: Number of best chains to include.
drop: Logical; whether to drop the list if only one parameter is selected.
...: Additional arguments passed to `estimate()`.

Returns

A named list of EAP estimates.

Method `MAP()`

Calculate Maximum A Posteriori (MAP) estimates.

Usage

RTMB_Fit_Base$MAP(
  pars = "parameters",
  chains = NULL,
  best_chains = NULL,
  type = c("marginal", "joint"),
  drop = TRUE,
  ...
)

Arguments

pars: Optional character vector of parameter names to extract.
chains: Numeric vector of chains to include.
best_chains: Number of best chains to include.
type: Character string; "marginal" or "joint" MAP.
drop: Logical; whether to drop the list if only one parameter is selected.
...: Additional arguments passed to `estimate()`.

Returns

A named list of MAP estimates.

Method `rotate()`

Rotate sampled parameters.

Usage

RTMB_Fit_Base$rotate(
  target,
  reference = NULL,
  linked = NULL,
  principal = FALSE,
  suffix = "rot",
  overwrite = TRUE
)

Arguments

target: Character string specifying the target variable to base the rotation on.
reference: Matrix to rotate towards. If NULL, the target's point estimate is used.
linked: Character vector of variable names to be rotated in the same direction.
principal: Logical; if TRUE and `reference` is NULL, the point estimate is first rotated to its principal axes and then used as the Procrustes reference. This is useful for MDU configurations.
suffix: Character suffix used for generated quantities. The default stores results as `<variable>_rot`.
overwrite: Logical; if TRUE, existing generated quantities with the same names are overwritten. If FALSE, a numeric suffix is appended.

Returns

The updated object invisibly.

Method `posterior_predict()`

Generate posterior predictive replications.

Usage

RTMB_Fit_Base$posterior_predict(
  code = NULL,
  variable = NULL,
  draws = 100L,
  seed = NULL,
  random = c("conditional", "population", "simulate"),
  chains = NULL,
  best_chains = NULL,
  observed = NULL
)

Arguments

code: Optional `rtmb_code()` object containing a `generate` block for custom models. The block should create and report one replicated outcome.
variable: Name of the replicated quantity returned by `code`. If omitted, `y_rep` or the only returned quantity is used.
draws: Number of predictive draws, or `NULL` to use every available posterior draw.
seed: Optional random seed.
random: Hierarchical prediction mode: `"conditional"` uses fitted random effects, `"population"` omits them, and `"simulate"` draws new random effects for the observed groups.
chains: Optional chains to use.
best_chains: Optional number of best chains to use.
observed: Optional observed outcome or its name in model data. This is stored as metadata when its length matches a prediction.

Returns

A matrix with predictive draws in rows and observations in columns.

Method `pp_check()`

Plot a posterior predictive check.

Usage

RTMB_Fit_Base$pp_check(
  type = c("auto", "dens", "bars"),
  stat = NULL,
  x = NULL,
  code = NULL,
  observed = NULL,
  variable = NULL,
  draws = 100L,
  seed = NULL,
  random = c("conditional", "population", "simulate"),
  chains = NULL,
  best_chains = NULL,
  plot = TRUE,
  ...
)

Arguments

type: Plot type: `"dens"`, `"bars"`, or `"auto"`. Automatic mode uses the observation likelihood's `_lpdf` or `_lpmf` implementation.
stat: Optional scalar statistic function, or its name. When supplied, the replicated statistic distribution is compared with its observed value.
x: Optional predictor used for a scatter-based check. Supply a column name from the model data or a vector with one value per observation.
code: Optional custom generate block; see `posterior_predict()`.
observed: Optional observed outcome or its name in model data.
variable: Name of the replicated quantity returned by `code`.
draws: Number of posterior predictive replications.
seed: Optional random seed.
random: Hierarchical prediction mode; see `posterior_predict()`.
chains: Optional chains to use.
best_chains: Optional number of best chains to use.
plot: Logical; draw the check immediately.
...: Graphical arguments passed to the plotting method.

Returns

An `rtmb_pp_check` object, invisibly.

Method `fa_rotate()`

Rotate factor loadings and optional factor scores.

Usage

RTMB_Fit_Base$fa_rotate(
  target = "L",
  linked = NULL,
  scores = NULL,
  rotate = "promax",
  ...,
  suffix = NULL,
  overwrite = TRUE
)

Arguments

target: Character string specifying the target variable to base the rotation on.
linked: Character vector of variable names to be rotated in the same direction.
scores: Character vector of variable names to be rotated as factor scores (inverse direction).
rotate: Character string specifying the rotation method.
...: Additional arguments passed to the rotation function.
suffix: Character suffix used for generated quantities. If NULL, the rotation method name is used.
overwrite: Logical; if TRUE, existing generated quantities with the same names are overwritten. If FALSE, a numeric suffix is appended.

Returns

The updated object invisibly.

Method `clone()`

The objects of this class are cloneable with this method.

Usage

RTMB_Fit_Base$clone(deep = FALSE)

Arguments

deep: Whether to make a deep clone.

Base class for RTMB Fit objects

Details

Public fields

Methods

Public methods

Method get_point_estimate()

Usage

Arguments

Returns

Method estimate()

Usage

Arguments

Returns

Method EAP()

Usage

Arguments

Returns

Method MAP()

Usage

Arguments

Returns

Method rotate()

Usage

Arguments

Returns

Method posterior_predict()

Usage

Arguments

Returns

Method pp_check()

Usage

Arguments

Returns

Method fa_rotate()

Usage

Arguments

Returns

Method clone()

Usage

Arguments

Method `get_point_estimate()`

Method `estimate()`

Method `EAP()`

Method `MAP()`

Method `rotate()`

Method `posterior_predict()`

Method `pp_check()`

Method `fa_rotate()`

Method `clone()`