Package 'adlaplaceHgp'

Conditional simulation for IWP/HIWP terms

Description

Draw conditional simulations of the Gaussian process model components, then summarize the resulting linear predictors and group-level effect curves.

Usage

cond_sim_iwp(
  fit,
  model_data,
  newx = NULL,
  n = 500,
  weights = NULL,
  probs = c(0.025, 0.5, 0.975),
  probs_envelope = c(0.1, 0.9)
)

Arguments

fit	Output of log_lik_laplace(..., deriv = TRUE).
model_data	Output of model_data().
newx	Optional list of prediction data frames, one per variable. When NULL, a default prediction grid is built from term knot ranges.
n	Number of conditional draws to simulate.
weights	Optional weights used to average group-level effects.
probs	Numeric vector of probabilities used when computing quantiles.
probs_envelope	Numeric vector of probabilities for envelopes.

Value

A list with components sim, quantiles, and envelope.

See Also

cond_sim_iwp_at, cond_sim_iwp_inputs

---

Conditional simulation for IWP/HIWP terms (flat arguments)

Description

Draw conditional simulations of IWP/HIWP/RSIWP model components from explicit inputs. Prefer cond_sim_iwp when you have a Laplace fit and model_data objects; use this function for custom pipelines.

Usage

cond_sim_iwp_at(
  terms,
  random_info,
  beta,
  gamma_mode,
  half_H_inv,
  newx = NULL,
  n = 500,
  weights = NULL,
  probs = c(0.025, 0.5, 0.975),
  probs_envelope = c(0.1, 0.9)
)

Arguments

terms	Named list of model term objects from model_data()$terms.
random_info	Data frame from model_data()$data$info$gamma with columns gamma_label, term, model, and by.
beta	Named numeric vector of fixed-effect MLEs, or a data frame with columns beta_label and mle.
gamma_mode	Named numeric vector of random-effect modes at the Laplace inner optimum, or a data frame with columns gamma_label and mode.
half_H_inv	Matrix $H^{-1/2}$ for inner random effects; from laplace_half_H_inv(laplace).
newx	Optional list of prediction data frames, one per variable. When NULL, a default grid is built from term knot ranges.
n	Number of conditional draws.
weights	Optional group weights (named vector or list by variable).
probs	Quantile levels for summaries.
probs_envelope	Envelope probability levels (requires GET).

Value

List with simulated curves, quantiles, and envelopes.

---

Build inputs for conditional IWP simulation

Description

Extracts flat arguments for cond_sim_iwp_at from the output of log_lik_laplace and model_data.

Usage

cond_sim_iwp_inputs(laplace, model_data)

Arguments

laplace	Output of log_lik_laplace(..., deriv = TRUE).
model_data	Output of model_data().

Value

A list with terms, random_info, beta, gamma_mode, and half_H_inv.

---

Hierarchical Integrated Wiener Process Model Term

Description

Creates and manages hierarchical integrated Wiener process (HIWP) model terms.

Usage

hiwp(
  x,
  p = 2,
  ref_value = 0,
  knots,
  by,
  init = NULL,
  lower = NULL,
  upper = NULL,
  parscale = NULL,
  boundary_is_random = TRUE,
  include_poly = TRUE,
  include_global = TRUE
)

## S4 method for signature 'hiwp'
design(term, data)

## S4 method for signature 'hiwp'
precision(term, data)

## S4 method for signature 'hiwp'
theta_info(term)

## S4 method for signature 'hiwp'
beta_info(term, data)

## S4 method for signature 'hiwp'
random_info(term, data)

Arguments

x	Variable name
p	Order of the integrated Wiener process (default: 2)
ref_value	Reference value for the basis
knots	Vector of knot locations
by	Grouping variable for hierarchical structure
init	Initial values for theta parameters
lower	Lower bounds for theta parameters
upper	Upper bounds for theta parameters
parscale	Parameter scales for optimization
boundary_is_random	Whether boundary should be treated as random
include_poly	Whether to include polynomial terms
include_global	Whether to include global component
term	An hiwp term object.
data	A data frame containing the variables used in the term.

Details

Hierarchical Integrated Wiener Process Model Term

Value

A list containing hiwp term object and optionally related terms

A design matrix for the HIWP term.

A precision matrix for the HIWP term.

A data frame containing theta parameter information for the HIWP term.

NULL (HIWP terms don't have beta parameters).

A data frame containing random effects information for the HIWP term.

Methods

The following methods are available for hiwp objects:

design(term, data)

Creates design matrix for HIWP term

precision(term, data)

Creates precision matrix for HIWP term

theta_info(term)

Extracts theta parameter information

beta_info(term, data)

Extracts beta parameter information

random_info(term, data)

Extracts random effects information

Examples

# Example usage:
# knots <- seq(0, 1, length.out = 5)
# hiwp_term <- hiwp(x = "age", knots = knots, by = "group")

---

Hierarchical Random Polynomial Term Constructor

Description

Creates and manages hierarchical random polynomial model terms.

Creates a hierarchical random polynomial model term.

Usage

hrpoly(
  x,
  p = 1,
  ref_value = 0,
  by,
  init = NULL,
  lower = NULL,
  upper = NULL,
  parscale = NULL
)

## S4 method for signature 'hrpoly'
design(term, data)

## S4 method for signature 'hrpoly'
precision(term, data)

## S4 method for signature 'hrpoly'
theta_info(term)

## S4 method for signature 'hrpoly'
beta_info(term, data)

## S4 method for signature 'hrpoly'
random_info(term, data)

Arguments

x	Variable name.
p	Polynomial degree (default: 1).
ref_value	Reference value for the polynomial.
by	Grouping variable for hierarchical structure.
init	Initial value for theta parameter.
lower	Lower bound for theta parameter.
upper	Upper bound for theta parameter.
parscale	Parameter scale for optimization.
term	An hrpoly term object.
data	A data frame containing the variables used in the term.

Value

An hrpoly term object.

A design matrix for the hrpoly term.

A precision matrix for the hrpoly term.

A data frame containing theta parameter information for the hrpoly term.

NULL (hrpoly terms don't have beta parameters).

A data frame containing random effects information for the hrpoly term.

Methods

The following methods are available for hrpoly objects:

design(term, data)

Creates design matrix for hrpoly term

precision(term, data)

Creates precision matrix for hrpoly term

theta_info(term)

Extracts theta parameter information

beta_info(term, data)

Extracts beta parameter information

random_info(term, data)

Extracts random effects information

Examples

# Create an hrpoly term for variable 'age' with grouping by 'site'
term <- hrpoly(x = "age", p = 2, ref_value = 0, by = "site")
str(term)
# Create sample data
set.seed(42)
dat <- data.frame(
  age = rep(seq(0, 80, by = 10), each = 3),
  site = rep(c("A", "B", "C"), times = 9)
)
by_group(term@by@term, dat)
# Compute the design matrix
design_mat <- design(term, dat)
print(design_mat)

---

Random Slope Fixed Polynomial Model Term

Description

Creates a fixed polynomial model term for random slope models.

Usage

rsfpoly(
  x,
  mult,
  p = 2,
  ref_value = 0,
  ref_mult = 0,
  init = NULL,
  lower = NULL,
  upper = NULL,
  parscale = NULL
)

## S4 method for signature 'rsfpoly'
design(term, data)

## S4 method for signature 'rsfpoly'
precision(term, data)

## S4 method for signature 'rsfpoly'
theta_info(term)

## S4 method for signature 'rsfpoly'
beta_info(term, data)

## S4 method for signature 'rsfpoly'
random_info(term, data)

Arguments

x	Variable name.
mult	Variable to multiply the polynomial by.
p	Polynomial degree (default: 2).
ref_value	Reference value for the polynomial.
ref_mult	Reference value for the covariate.
init	Initial values for beta parameters.
lower	Lower bounds for beta parameters.
upper	Upper bounds for beta parameters.
parscale	Parameter scales for optimization.
term	A rsfpoly term object
data	A data frame containing the term variables

Value

A rsfpoly term object.

A design matrix for the random slope fixed polynomial term, or NULL if p.order is 0

NULL (fixed effects don't have precision matrices)

NULL (fixed effects don't have theta parameters)

A data frame containing beta parameter information for the random slope fixed polynomial term

NULL (fixed effects don't have random effects information)

Functions

• design(rsfpoly): Creates design matrix for rsfpoly term

• precision(rsfpoly): Creates precision matrix for rsfpoly term

• theta_info(rsfpoly): Extracts theta parameter information for rsfpoly term

• beta_info(rsfpoly): Extracts beta parameter information for rsfpoly term

• random_info(rsfpoly): Extracts random effects information for rsfpoly term

Methods

The following methods are available for rsfpoly objects:

design(term, data)

Creates design matrix for rsfpoly term

precision(term, data)

Creates precision matrix for rsfpoly term

theta_info(term)

Extracts theta parameter information

beta_info(term, data)

Extracts beta parameter information

random_info(term, data)

Extracts random effects information

---

independent Random Slope

Description

Creates an iid random slope model term

Usage

rsiid(
  x,
  mult,
  ref_mult = 0,
  init = NULL,
  lower = NULL,
  upper = NULL,
  parscale = NULL
)

## S4 method for signature 'rsiid'
design(term, data)

## S4 method for signature 'rsiid'
precision(term, data)

## S4 method for signature 'rsiid'
theta_info(term)

## S4 method for signature 'rsiid'
beta_info(term, data)

## S4 method for signature 'rsiid'
random_info(term, data)

Arguments

x	Variable name.
mult	Variable to multiply the polynomial by.
ref_mult	Reference value for the covariate.
init	Initial values for beta parameters.
lower	Lower bounds for beta parameters.
upper	Upper bounds for beta parameters.
parscale	Parameter scales for optimization.
term	A rsiid term object
data	A data frame containing the term variables

Value

A rsiid term object.

A design matrix for the random slope polynomial term, or NULL if p.order is 0

A precision matrix for the random slope polynomial term

A data frame containing theta parameter information for the random slope term

NULL (random slope polynomial terms don't have beta parameters)

A data frame containing random effects information for the random slope polynomial term

Functions

• design(rsiid): Creates design matrix for rsiid term

• precision(rsiid): Creates precision matrix for rsiid term

• theta_info(rsiid): Extracts theta parameter information for rsiid term

• beta_info(rsiid): Extracts beta parameter information for rsiid term

• random_info(rsiid): Extracts random effects information for rsiid term

Methods

The following methods are available for rsiid objects:

design(term, data)

Creates design matrix for rsiid term

precision(term, data)

Creates precision matrix for rsiid term

theta_info(term)

Extracts theta parameter information

beta_info(term, data)

Extracts beta parameter information

random_info(term, data)

Extracts random effects information

---

Random Slope Integrated Wiener Process Term Constructor

Description

Creates a random slope integrated Wiener process (RSIWP) model term.

Usage

rsiwp(
  x,
  mult,
  p = 2,
  ref_value = 0,
  ref_mult = 0,
  knots,
  init = NULL,
  lower = NULL,
  upper = NULL,
  parscale = NULL,
  boundary_is_random = TRUE,
  include_poly = TRUE,
  include_linear = TRUE
)

Arguments

x	Variable name.
mult	Variable to multiply the IWP by.
p	Order of the integrated Wiener process (default: 2).
ref_value	Reference value for the basis.
ref_mult	Reference value for the covariate.
knots	Vector of knot locations.
init	Initial values for theta parameters.
lower	Lower bounds for theta parameters.
upper	Upper bounds for theta parameters.
parscale	Parameter scales for optimization.
boundary_is_random	Whether boundary should be treated as random.
include_poly	Whether to include polynomial terms.
include_linear	Whether to include a linear (fixed) term in mult.

Value

A list containing the rsiwp term object and optionally polynomial terms.

---

Random Slope Integrated Wiener Process Term

Description

Creates and manages random slope integrated Wiener process (RSIWP) model terms.

Usage

## S4 method for signature 'rsiwp'
design(term, data)

## S4 method for signature 'rsiwp'
precision(term, data)

## S4 method for signature 'rsiwp'
theta_info(term)

## S4 method for signature 'rsiwp'
beta_info(term, data)

## S4 method for signature 'rsiwp'
random_info(term, data)

Arguments

term	An rsiwp term object
data	A data frame containing the term variable

Functions

• design(rsiwp): Creates design matrix for RSIWP term

• precision(rsiwp): Creates precision matrix for RSIWP term

• theta_info(rsiwp): Extracts theta parameter information for RSIWP term

• beta_info(rsiwp): Extracts beta parameter information for RSIWP term

• random_info(rsiwp): Extracts random effects information for RSIWP term

Methods

The following methods are available for rsiwp objects:

design(term, data)

Creates design matrix for RSIWP term

precision(term, data)

Creates precision matrix for RSIWP term

theta_info(term)

Extracts theta parameter information

beta_info(term, data)

Extracts beta parameter information

random_info(term, data)

Extracts random effects information

---

Random Polynomial Model for Random Slope

Description

Creates a random polynomial model term for use in a random slope model.

Usage

rsrpoly(x, mult, p = 2, ref_value = 0, ref_mult = 0, sd = Inf)

## S4 method for signature 'rsrpoly'
design(term, data)

## S4 method for signature 'rsrpoly'
precision(term, data)

## S4 method for signature 'rsrpoly'
theta_info(term)

## S4 method for signature 'rsrpoly'
beta_info(term, data)

## S4 method for signature 'rsrpoly'
random_info(term, data)

Arguments

x	Variable name.
mult	Variable to multiply the polynomial by.
p	Polynomial degree (default: 2).
ref_value	Reference value for the polynomial.
ref_mult	Reference value for the covariate.
sd	Standard deviation for random effects.
term	A rsrpoly term object
data	A data frame containing the term variables

Value

A rsrpoly term object.

A design matrix for the random slope polynomial term, or NULL if p.order is 0

A precision matrix for the random slope polynomial term

NULL (random slope polynomial terms don't have theta parameters)

NULL (random slope polynomial terms don't have beta parameters)

A data frame containing random effects information for the random slope polynomial term

Functions

• design(rsrpoly): Creates design matrix for rsrpoly term

• precision(rsrpoly): Creates precision matrix for rsrpoly term

• theta_info(rsrpoly): Extracts theta parameter information for rsrpoly term

• beta_info(rsrpoly): Extracts beta parameter information for rsrpoly term

• random_info(rsrpoly): Extracts random effects information for rsrpoly term

Methods

The following methods are available for rsrpoly objects:

design(term, data)

Creates design matrix for rsrpoly term

precision(term, data)

Creates precision matrix for rsrpoly term

theta_info(term)

Extracts theta parameter information

beta_info(term, data)

Extracts beta parameter information

random_info(term, data)

Extracts random effects information

---