R/create_empty_list.R
In christian-gische/causalSEM: Analyze Causal Effects Using Interventional Distributions
Defines functions
create_empty_list
Documented in
create_empty_list
## Changelog:
# CG 0.0.12 2023-02-21: changes in preamble and comments
# MH 0.0.11 2022-03-17: removed "\code" from "\code{\tabular{lll}..."
# solve WARNING in package checking
# CG 0.0.10 2022-01-13: changed structure of internal_list
# cleaned up code (documentation, 80 char per line)
# changed dot-case to snake-case
# MH 0.0.9 2021-11-30: changed defaults of interventional_distribution$
# density_function$values to empty list
# changed "pdf" to "pdfs"
# updated documentation accordingly
# MH 0.0.8 2021-11-22: renamed from make_empty_list to create_empty_list
# MH 0.0.7 2021-11-22: moments$variance_matrix changed to covariance_matrix
# CG 0.0.6 2021-11-22: create slots for select outcome matrix
# CG 0.0.5 2021-11-09: create slots for eliminiation, duplication
# and commutation matrix
# CG 0.0.4 2021-09-24: create slot 'interventional_distribution'
# update Documentation
# MA 0.0.3 2021-09-09: added 'derivative' slot to 'C' and 'Psi'
# CG 0.0.2 2021-09-02: create slot 'param'
# MH 0.0.1 2021-09-01: initial programming
## Documentation
#' @title Create Empty Internal List
#' @description Creates an empty internal list of predefined structure. The
#' list will subsequently be filled while the function
#' \code{\link{intervention_effect}} is running.
#' @param verbose Integer number describing verbosity of console output.
#' 0...no output (default); 1...user messages; 2...debugging-relevant messages.
#' @return An empty internal list with the following structure:\cr
#' \tabular{lll}{
#' List of 6\cr
#' \code{$fitted_object}: List \tab \tab Object containing fitted
#' model. \cr
#' \code{$fitted_object_class}: chr(0) \tab \tab Class of fitted model. \cr
#' \code{$info_model}: List of 6 \tab \tab \cr
#' \code{..$n_obs}: int(0) \tab \tab Number of observations.\cr
#' \code{..$n_ov}: int(0) \tab \tab Number of manifest variables.\cr
#' \code{..$var_names}: chr(0) \tab \tab Names of observed variables.\cr
#' \code{..$C}: List of 3 \tab \tab \cr
#' \code{.. ..$values}: num[0, 0] \tab \tab Parameter values of matrix of
#' structural coefficients.\cr
#' \code{.. ..$labels}: chr[0, 0] \tab \tab Parameter labels of matrix of
#' structural coefficients.\cr
#' \code{.. ..$derivative}: chr[0, 0] \tab \tab Partial derivative of the
#' vectorized C matrix with \cr
#' \tab \tab respect to the parameters.\cr
#' \code{..$ Psi}: List of 3 \tab \tab \cr
#' \code{.. ..$values}: num[0, 0] \tab \tab Parameter values of model
#' implied covariance matrix.\cr
#' \code{.. ..$labels}: chr[0, 0] \tab \tab Parameter labels of model
#' implied covariance matrix\cr
#' \code{.. ..$derivative}: chr[0, 0] \tab \tab Partial derivative of the
#' vectorized Psi matrix with\cr
#' \tab \tab respect to the parameters.\cr
#' \code{..$param}: List of 5 \tab \tab \cr
#' \code{.. ..$n_par}: int(0) \tab \tab Total number of estimated
#' parameters.\cr
#' \code{.. ..$n_par_unique}: int(0) \tab \tab Number of distinct and
#' functionally unrelated parameters.\cr
#' \code{.. ..$labels_par_unique}: chr(0) \tab \tab Labels of distinct and
#' functionally unrelated parameters.\cr
#' \code{.. ..$values_par_unique}: num(0) \tab \tab Values of distinct
#' and functionally unrelated parameters.\cr
#' \code{.. ..$varcov_par_unique}: num[0, 0] \tab \tab Covariance
#' matrix of the estimator of distinct\cr
#' \tab \tab and functionally unrelated parameters.\cr
#' \code{$info_interventions}: List of 8 \tab \tab \cr
#' \code{..$n_intervention}: int(0) \tab \tab Number of interventional
#' variables.\cr
#' \code{..$intervention_names}: chr(0) \tab \tab Names of interventional
#' variables.\cr
#' \code{..$n_outcome}: int(0) \tab \tab Number of outcome variables.\cr
#' \code{..$outcome_names}: chr(0) \tab \tab Names of outcome variables.\cr
#' \code{..$intervention_levels}: num(0) \tab \tab Interventional levels.\cr
#' \code{..$effect_type}: chr(0) \tab \tab Features of the interventional
#' distribution to be analyzed.\cr
#' \code{..$lower_bound}: num(0) \tab \tab Lower bound of critical range
#' of outcome variables.\cr
#' \code{..$upper_bound}: num(0) \tab \tab Upper bound of critical range
#' of outcome variables.\cr
#' \code{$matrices_constant}: List of 7 \tab \tab \cr
#' \code{.. .. $select_intervention}: num[0, 0] \tab \tab Selection matrix
#' for entries that are intervened on. \cr
#' \code{.. .. $select_non_intervention}: num[0, 0] \tab \tab Selection
#' matrix for entries that are NOT intervened on. \cr
#' \code{.. .. $select_outcome}: num[0, 0] \tab \tab Selection
#' matrix for outcomes of interest. \cr
#' \code{.. .. $eliminate_intervention}: num[0, 0] \tab \tab Matrix that
#' replaces entries that are intervened on by zero. \cr
#' \code{.. .. $duplication_matrix}: num[0, 0] \tab \tab Maps vech(A)
#' onto vec(A) for symmetric A. \cr
#' \code{.. .. $elimination_matrix}: num[0, 0] \tab \tab Maps vec(A)
#' onto vech(A). \cr
#' \code{.. .. $commutation_matrix}: num[0, 0] \tab \tab Maps vech(A)
#' onto vec(A'). \cr
#' \code{$interventional_distribution} : List of 4 \tab \tab \cr
#' \code{..$means}: List of 4 \tab \tab \cr
#' \code{.. .. $values}: num(0) \tab \tab Mean vector of the
#' interventional distribution. \cr
#' \code{.. .. $jacobian}: num[0, 0] \tab \tab Jacobian of interventinoal
#' mean vector. \cr
#' \code{.. .. $ase}: num(0) \tab \tab Asymptotic standard errors of
#' interventional means. \cr
#' \code{.. .. $z_values}: num[0, 0] \tab \tab z-values of interventional
#' means.\cr
#' \code{..$covariance_matrix}: List of 4 \tab \tab \cr
#' \code{.. .. $values}: num(0) \tab \tab Covariance matrix of the
#' interventional distribution. \cr
#' \code{.. .. $jacobian}: num[0, 0] \tab \tab Jacobian of VECTORIZED
#' interventional covariance matrix. \cr
#' \code{.. .. $ase}: num(0) \tab \tab Asymptotic standard errors of
#' interventional (co)variances. \cr
#' \code{.. .. $z_values}: num[0, 0] \tab \tab z-values of interventional
#' interventional (co)variances. \cr
#' \code{..$density_function}: List of 4 \tab \tab \cr
#' \code{.. .. $values}: num(0) \tab \tab Probability density function
#' (pdf) of the\cr
#' \tab \tab interventional distribution. \cr
#' \code{.. .. $jacobian}: num[0, 0] \tab \tab Jacobian of the
#' interventional pdf.\cr
#' \code{.. .. $ase}: num(0) \tab \tab asymptotic standard errors of
#' interventional pdf. \cr
#' \code{.. .. $z_values}: num[0, 0] \tab \tab z-values of interventional
#' pdf. \cr
#' \code{..$probabilities}: List of 4 \tab \tab \cr
#' \code{.. .. $values}: num(0) \tab \tab Probabilities of
#' interventional events. \cr
#' \code{.. .. $jacobian}: num[0, 0] \tab \tab Jacobian of interventional
#' probabilities. \cr
#' \code{.. .. $ase}: num(0) \tab \tab Asymptotic standard error of
#' interventional probabilities. \cr
#' \code{.. .. $z_values}: num[0, 0] \tab \tab z-value of the
#' interventional probabilities. \cr
#' \code{$control}: List of 1 \tab \tab \cr
#' \code{..$verbose}: num(0) \tab \tab Verbosity of console output.\cr
#' \code{$tables}: List of 2 \tab \tab \cr
#' \code{..$interventional_means}: data.frame \tab \tab Information for
#' \code{\link{summary.causalSEM}} \cr
#' \tab \tab and \code{\link{print.causalSEM}} functions.\cr
#' \code{..$interventional_means}: data.frame \tab \tab Information for
#' \code{\link{summary.causalSEM}}\cr
#' \tab \tab and \code{\link{print.causalSEM}}
#' functions.\cr
#' }
#'
#' @references Gische, C., Voelkle, M.C. (2022) Beyond the Mean: A Flexible
#' Framework for Studying Causal Effects Using Linear Models. Psychometrika 87,
#' 868–901. https://doi.org/10.1007/s11336-021-09811-z
## Function definition
create_empty_list <- function( verbose = NULL ){
# function name
fun.name <- "create_empty_list"
# function version
fun.version <- "0.0.12 2023-02-21"
# function name+version
fun.name.version <- paste0( fun.name, " (", fun.version, ")" )
# check verbose argument
verbose <- handle_verbose_argument( verbose )
# console output
if( verbose >= 2 ) cat( paste0( "start of function ", fun.name.version, " ",
Sys.time(), "\n" ) )
# internal list
internal_list <- list(
# fitted object
"fitted_object" = NULL,
# class of fitted object
"fitted_object_class" = NULL,
# model info
info_model = list(
# number of observations
# a single number, normally an integer
"n_obs" = as.integer(0),
# number of manifest variables
# a single number, normally an integer
"n_ov" = as.integer(0),
# names of observed variables
# character vector of length(n_ov)
# general: using all-y LISREL notation
# here: using the notation from Gische and Voelkle (2022).
"var_names" = character(0),
# C list
C = list(
# parameter values of matrix of structural coefficients
# numeric matrix of dimension n_ov * n_ov
"values" = matrix(numeric(0))[-1,-1],
# parameter names of matrix of structural coefficients
# character matrix of dimension n_ov * n_ov
# same labels display equality constraints
# general: using all-y LISREL notation
# here: using the notation from Gische and Voelkle (2022)
"labels" = matrix(character(0))[-1,-1],
# Partial derivative of the vectorized C matrix with respect
# to the parameters
"derivative" = matrix(numeric(0))[-1,-1]
), # end of C list
# Psi list
Psi = list(
# parameter values of model implied covariance matrix
# numeric matrix of dimension n_ov * n_ov
"values" = matrix(numeric(0))[-1,-1],
# parameter names of model implied covariance matrix
# character matrix of dimension n_ov * n_ov
# same labels display equality constraints
# general: using all-y LISREL notation
# here: using the notation from Gische and Voelkle (2022)
"labels" = matrix(character(0))[-1,-1],
# Partial derivative of the vectorized Psi matrix with respect
# to the parameters
"derivative" = matrix(numeric(0))[-1,-1]
), # end of Psi list
param = list(
# total number of estimated parameters
# normally an integer
# equals the total number of non-NA entries in
# structural_coeff_names and covariance_names
# does NOT take into account other specified equality constraints
"n_par" = integer(0),
# number of distinct and functionally unrelated parameters
# normally an integer
"n_par_unique" = as.integer(0),
# names of distinct and functionally unrelated parameters
# in the order they appear rowwise in the matrices
# structural_coeff_names and covariance_names
# character vector of length(n_par_unique)
"labels_par_unique" = character(0),
# parameter values (estimates) of distinct and functionally unrelated
# parameters
# numeric vector of length(n_par_unique)
"values_par_unique" = numeric(0),
# variance-covariance matrix of estimator of vector
# of distinct and functionally unrelated parameters
# numeric matrix of dimension n_par_unique * n_par_unique
"varcov_par_unique" = matrix(numeric(0))[-1,-1]
) # end of param list
), # end of info_model list
# interventions
info_interventions = list(
# number of interventional variables
# normally an integer
# default: zero
"n_intervention" = as.integer(0),
# names of interventional variables
# character vector of length n_intervention
# default: empty vector (no intervention)
"intervention_names" = character(0),
# number of outcome variables
# normally an integer
# default: n_ov - n_intervention
"n_outcome" = as.integer(0),
# names of outcome variables
# character vector of length(n_outcome)
# default: all non-interventional variables
"outcome_names" = character(0),
# interventional level
# numeric vector of length(n_intervention)
# default: vector of ones
"intervention_levels" = numeric(0),
# parts of the interventional distribution the user is interested in
# character vector
# if "probability" is specified, user specified values of
# lower_bound and upper_bound required
# default: "mean"
# possible values: c("mean", "variance", "density", "probability")
"effect_type" = character(0),
# lower bound of critical range of outcome variables
# numeric vector
"lower_bounds" = numeric(0),
# upper bound of critical range of outcome variables
# numeric vector
"upper_bounds" = numeric(0)
), # end of info_interventions list
# constant matrices for computations
constant_matrices = list(
# selection matrix for entries that are intervened on
# see Definition 1 point 3 in Gische and Voelkle (2022)
# numeric matrix of dimension n_ov x n_intervention
"select_intervention" = matrix(numeric(0))[-1,-1],
# selection matrix for entries that are NOT intervened on
# see Definition 1 point 3 in Gische and Voelkle (2022)
# numeric matrix of dimension n_ov x (n_ov-n_intervention)
"select_non_intervention" = matrix(numeric(0))[-1,-1],
# matrix that replaces entries that are intervened on by zero
# see Definition 1 point 4 in Gische and Voelkle (2022)
# numeric matrix of dimension n_ov x n_ov
"eliminate_intervention" = matrix(numeric(0))[-1,-1],
# selection matrix for outcome variables of interest
# numeric matrix of dimension n_ov x (number of outcome variables
# of interest)
"select_outcome" = matrix(numeric(0))[-1,-1],
# zero-one matrix that maps vech(A) onto vec(A)
# for symmetric A
# numeric matrix of dimension (n_ov^2) x (1/2*n_ov*(n_ov+1))
"duplication_matrix" = matrix(numeric(0))[-1,-1],
# zero-one matrix that eliminates from vec(A)
# the supradiagonal elements of A. In other words,
# it maps vec(A) onto vech(A).
# numeric matrix of dimension (1/2*n_ov*(n_ov+1)) x (n_ov^2)
"elimination_matrix" = matrix(numeric(0))[-1,-1],
# zero-one matrix that transforms vec(A) into vec(A')
# numeric matrix of dimension (n_ov^2) x (n_ov^2)
"commutation_matrix" = matrix(numeric(0))[-1,-1]
), # end of matrices_constant list
# interventional distribution
interventional_distribution = list(
means = list(
# mean vector of the interventional distribution
# see Equation (6a) in Gische and Voelkle (2022)
# numeric vector of length n_ov
"values" = numeric(0),
# jacobian of the mean vector
# numeric matrix of dimension n_ov x n_par_unique
"jacobian" = matrix(numeric(0))[-1,-1],
# asymptotic covariance matrix of the interventional mean vector
# numeric matrix of dimension n_outcomes x n_outcomes
"acov" = matrix(numeric(0))[-1,-1],
# asymptotic standard errors of the interventional means
# numeric vector of length n_ov
"ase" = numeric(0),
# z-values of the interventional means
# numeric vector of length n_ov
"z_values" = numeric(0)
), # end of means
covariance_matrix = list(
# covariance matrix of the interventional distribution
# see Equation (6b) in Gische and Voelkle (2022)
# numeric matrix of dimension n_ov x n_ov
"values" = matrix(numeric(0))[-1,-1],
# jacobian of the VECTORIZED variance-covariance matrix
# numeric matrix of dimension n_ov^2 x n_par_unique
"jacobian" = matrix(numeric(0))[-1,-1],
# asymptotic covariance matrix of the VECTORIZED covariance matrix of the
# interventional distribution
# numeric matrix of dimension n_outcomes^2 x n_outcomes^2
"acov" = matrix(numeric(0))[-1,-1],
# asymptotic standard errors of the interventional VARIANCES
# (i.e., the diagonal entries of the interventional covariance matrix)
# numeric vector of length n n_outcomes
"ase" = matrix(numeric(0))[-1,-1],
# z-values of the the interventional VARIANCES
# (i.e., the diagonal entries of the interventional covariance matrix)
"z_values" = matrix(numeric(0))[-1,-1]
), # end of covariance_matrix
density_function = list(
# probability density function (pdf) of the interventional distribution
# see Equation (9) in Gische and Voelkle (2022)
# numeric matrix of with two columns
# 1 column: grid that captures the range of the outcome variable
# 2 column: values of the pdf
"values" = matrix(numeric(0))[-1,-1],
# jacobian of the density function (pdf) of the interventional
# distribution
# numeric matrix of with two columns
# 1 column: grid that captures the range of the outcome variable
# 2 column: values of the jacobian
"jacobian" = matrix(numeric(0))[-1,-1],
# asymptotic standard errors (ase) of the probability density function
# (pdf) # of the interventional distribution
# numeric matrix of with two columns
# 1 column: grid that captures the range of the outcome variable
# 2 column: values of the ase
"ase" = matrix(numeric(0))[-1,-1],
# z-values of the probability density function (pdf)
# of the interventional distribution
# 1 column: grid that captures the range of the outcome variable
# 2 column: values of the z-values
"z_values" = matrix(numeric(0))[-1,-1]
), # end of density
probabilities = list(
# probabilities of interventional events
# see Equation (10) in Gische and Voelkle (2022)
# numeric vector of length n_outcomes
"values" = numeric(0),
# jacobian of the probability of a UNIVARIATE interventional event
# if n_outcomes > 1, the computation is done separately for each
# outcome variable
# numeric vector of length n_outcomes
"jacobian" = matrix(numeric(0))[-1,-1],
# asymptotic covariance of the probability of a UNIVARIATE
# interventional event
# if n_outcomes > 1, the computation is done separately for each
# outcome variable
# numeric vector of length n_outcomes
"acov" = matrix(numeric(0))[-1,-1],
# asymptotic standard error (ase) of the probability of a UNIVARIATE
# interventional event
# if n_outcomes > 1, the computation is done separately for each
# outcome variable
# numeric vector of length n_outcomes
"ase" = numeric(0),
# asymptotic z-value of the probability of a UNIVARIATE
# interventional event
# if n_outcomes > 1, the computation is done separately for each
# outcome variable
# numeric vector of length n_outcomes
"z_values" = numeric(0)
) # end of probability
), # end of interventional distribution
# control list
control = list(
# verbosity of console output
# a number, 0...no output (default), 1...user messages,
# 2...debugging-relevant messages
"verbose" = verbose
) # end of control list
) # end of internal list
# console output
if( verbose >= 2 ) cat( paste0( " end of function ", fun.name.version, " ",
Sys.time(), "\n" ) )
# return internal list
return( internal_list )
}
## test/development
# source( "c:/Users/martin/Dropbox/68_causalSEM/04_martinhecht/R/
# handle_verbose_argument.R" )
# internal_list <- create_empty_list()
# str( internal_list )
christian-gische/causalSEM documentation built on April 26, 2023, 10:36 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions create_empty_list
Documented in create_empty_list
## Changelog:
# CG 0.0.12 2023-02-21: changes in preamble and comments
# MH 0.0.11 2022-03-17: removed "\code" from "\code{\tabular{lll}..."
# solve WARNING in package checking
# CG 0.0.10 2022-01-13: changed structure of internal_list
# cleaned up code (documentation, 80 char per line)
# changed dot-case to snake-case
# MH 0.0.9 2021-11-30: changed defaults of interventional_distribution$
# density_function$values to empty list
# changed "pdf" to "pdfs"
# updated documentation accordingly
# MH 0.0.8 2021-11-22: renamed from make_empty_list to create_empty_list
# MH 0.0.7 2021-11-22: moments$variance_matrix changed to covariance_matrix
# CG 0.0.6 2021-11-22: create slots for select outcome matrix
# CG 0.0.5 2021-11-09: create slots for eliminiation, duplication
# and commutation matrix
# CG 0.0.4 2021-09-24: create slot 'interventional_distribution'
# update Documentation
# MA 0.0.3 2021-09-09: added 'derivative' slot to 'C' and 'Psi'
# CG 0.0.2 2021-09-02: create slot 'param'
# MH 0.0.1 2021-09-01: initial programming
## Documentation
#' @title Create Empty Internal List
#' @description Creates an empty internal list of predefined structure. The
#' list will subsequently be filled while the function
#' \code{\link{intervention_effect}} is running.
#' @param verbose Integer number describing verbosity of console output.
#' 0...no output (default); 1...user messages; 2...debugging-relevant messages.
#' @return An empty internal list with the following structure:\cr
#' \tabular{lll}{
#' List of 6\cr
#' \code{$fitted_object}: List \tab \tab Object containing fitted
#' model. \cr
#' \code{$fitted_object_class}: chr(0) \tab \tab Class of fitted model. \cr
#' \code{$info_model}: List of 6 \tab \tab \cr
#' \code{..$n_obs}: int(0) \tab \tab Number of observations.\cr
#' \code{..$n_ov}: int(0) \tab \tab Number of manifest variables.\cr
#' \code{..$var_names}: chr(0) \tab \tab Names of observed variables.\cr
#' \code{..$C}: List of 3 \tab \tab \cr
#' \code{.. ..$values}: num[0, 0] \tab \tab Parameter values of matrix of
#' structural coefficients.\cr
#' \code{.. ..$labels}: chr[0, 0] \tab \tab Parameter labels of matrix of
#' structural coefficients.\cr
#' \code{.. ..$derivative}: chr[0, 0] \tab \tab Partial derivative of the
#' vectorized C matrix with \cr
#' \tab \tab respect to the parameters.\cr
#' \code{..$ Psi}: List of 3 \tab \tab \cr
#' \code{.. ..$values}: num[0, 0] \tab \tab Parameter values of model
#' implied covariance matrix.\cr
#' \code{.. ..$labels}: chr[0, 0] \tab \tab Parameter labels of model
#' implied covariance matrix\cr
#' \code{.. ..$derivative}: chr[0, 0] \tab \tab Partial derivative of the
#' vectorized Psi matrix with\cr
#' \tab \tab respect to the parameters.\cr
#' \code{..$param}: List of 5 \tab \tab \cr
#' \code{.. ..$n_par}: int(0) \tab \tab Total number of estimated
#' parameters.\cr
#' \code{.. ..$n_par_unique}: int(0) \tab \tab Number of distinct and
#' functionally unrelated parameters.\cr
#' \code{.. ..$labels_par_unique}: chr(0) \tab \tab Labels of distinct and
#' functionally unrelated parameters.\cr
#' \code{.. ..$values_par_unique}: num(0) \tab \tab Values of distinct
#' and functionally unrelated parameters.\cr
#' \code{.. ..$varcov_par_unique}: num[0, 0] \tab \tab Covariance
#' matrix of the estimator of distinct\cr
#' \tab \tab and functionally unrelated parameters.\cr
#' \code{$info_interventions}: List of 8 \tab \tab \cr
#' \code{..$n_intervention}: int(0) \tab \tab Number of interventional
#' variables.\cr
#' \code{..$intervention_names}: chr(0) \tab \tab Names of interventional
#' variables.\cr
#' \code{..$n_outcome}: int(0) \tab \tab Number of outcome variables.\cr
#' \code{..$outcome_names}: chr(0) \tab \tab Names of outcome variables.\cr
#' \code{..$intervention_levels}: num(0) \tab \tab Interventional levels.\cr
#' \code{..$effect_type}: chr(0) \tab \tab Features of the interventional
#' distribution to be analyzed.\cr
#' \code{..$lower_bound}: num(0) \tab \tab Lower bound of critical range
#' of outcome variables.\cr
#' \code{..$upper_bound}: num(0) \tab \tab Upper bound of critical range
#' of outcome variables.\cr
#' \code{$matrices_constant}: List of 7 \tab \tab \cr
#' \code{.. .. $select_intervention}: num[0, 0] \tab \tab Selection matrix
#' for entries that are intervened on. \cr
#' \code{.. .. $select_non_intervention}: num[0, 0] \tab \tab Selection
#' matrix for entries that are NOT intervened on. \cr
#' \code{.. .. $select_outcome}: num[0, 0] \tab \tab Selection
#' matrix for outcomes of interest. \cr
#' \code{.. .. $eliminate_intervention}: num[0, 0] \tab \tab Matrix that
#' replaces entries that are intervened on by zero. \cr
#' \code{.. .. $duplication_matrix}: num[0, 0] \tab \tab Maps vech(A)
#' onto vec(A) for symmetric A. \cr
#' \code{.. .. $elimination_matrix}: num[0, 0] \tab \tab Maps vec(A)
#' onto vech(A). \cr
#' \code{.. .. $commutation_matrix}: num[0, 0] \tab \tab Maps vech(A)
#' onto vec(A'). \cr
#' \code{$interventional_distribution} : List of 4 \tab \tab \cr
#' \code{..$means}: List of 4 \tab \tab \cr
#' \code{.. .. $values}: num(0) \tab \tab Mean vector of the
#' interventional distribution. \cr
#' \code{.. .. $jacobian}: num[0, 0] \tab \tab Jacobian of interventinoal
#' mean vector. \cr
#' \code{.. .. $ase}: num(0) \tab \tab Asymptotic standard errors of
#' interventional means. \cr
#' \code{.. .. $z_values}: num[0, 0] \tab \tab z-values of interventional
#' means.\cr
#' \code{..$covariance_matrix}: List of 4 \tab \tab \cr
#' \code{.. .. $values}: num(0) \tab \tab Covariance matrix of the
#' interventional distribution. \cr
#' \code{.. .. $jacobian}: num[0, 0] \tab \tab Jacobian of VECTORIZED
#' interventional covariance matrix. \cr
#' \code{.. .. $ase}: num(0) \tab \tab Asymptotic standard errors of
#' interventional (co)variances. \cr
#' \code{.. .. $z_values}: num[0, 0] \tab \tab z-values of interventional
#' interventional (co)variances. \cr
#' \code{..$density_function}: List of 4 \tab \tab \cr
#' \code{.. .. $values}: num(0) \tab \tab Probability density function
#' (pdf) of the\cr
#' \tab \tab interventional distribution. \cr
#' \code{.. .. $jacobian}: num[0, 0] \tab \tab Jacobian of the
#' interventional pdf.\cr
#' \code{.. .. $ase}: num(0) \tab \tab asymptotic standard errors of
#' interventional pdf. \cr
#' \code{.. .. $z_values}: num[0, 0] \tab \tab z-values of interventional
#' pdf. \cr
#' \code{..$probabilities}: List of 4 \tab \tab \cr
#' \code{.. .. $values}: num(0) \tab \tab Probabilities of
#' interventional events. \cr
#' \code{.. .. $jacobian}: num[0, 0] \tab \tab Jacobian of interventional
#' probabilities. \cr
#' \code{.. .. $ase}: num(0) \tab \tab Asymptotic standard error of
#' interventional probabilities. \cr
#' \code{.. .. $z_values}: num[0, 0] \tab \tab z-value of the
#' interventional probabilities. \cr
#' \code{$control}: List of 1 \tab \tab \cr
#' \code{..$verbose}: num(0) \tab \tab Verbosity of console output.\cr
#' \code{$tables}: List of 2 \tab \tab \cr
#' \code{..$interventional_means}: data.frame \tab \tab Information for
#' \code{\link{summary.causalSEM}} \cr
#' \tab \tab and \code{\link{print.causalSEM}} functions.\cr
#' \code{..$interventional_means}: data.frame \tab \tab Information for
#' \code{\link{summary.causalSEM}}\cr
#' \tab \tab and \code{\link{print.causalSEM}}
#' functions.\cr
#' }
#'
#' @references Gische, C., Voelkle, M.C. (2022) Beyond the Mean: A Flexible
#' Framework for Studying Causal Effects Using Linear Models. Psychometrika 87,
#' 868–901. https://doi.org/10.1007/s11336-021-09811-z
## Function definition
create_empty_list <- function( verbose = NULL ){
# function name
fun.name <- "create_empty_list"
# function version
fun.version <- "0.0.12 2023-02-21"
# function name+version
fun.name.version <- paste0( fun.name, " (", fun.version, ")" )
# check verbose argument
verbose <- handle_verbose_argument( verbose )
# console output
if( verbose >= 2 ) cat( paste0( "start of function ", fun.name.version, " ",
Sys.time(), "\n" ) )
# internal list
internal_list <- list(
# fitted object
"fitted_object" = NULL,
# class of fitted object
"fitted_object_class" = NULL,
# model info
info_model = list(
# number of observations
# a single number, normally an integer
"n_obs" = as.integer(0),
# number of manifest variables
# a single number, normally an integer
"n_ov" = as.integer(0),
# names of observed variables
# character vector of length(n_ov)
# general: using all-y LISREL notation
# here: using the notation from Gische and Voelkle (2022).
"var_names" = character(0),
# C list
C = list(
# parameter values of matrix of structural coefficients
# numeric matrix of dimension n_ov * n_ov
"values" = matrix(numeric(0))[-1,-1],
# parameter names of matrix of structural coefficients
# character matrix of dimension n_ov * n_ov
# same labels display equality constraints
# general: using all-y LISREL notation
# here: using the notation from Gische and Voelkle (2022)
"labels" = matrix(character(0))[-1,-1],
# Partial derivative of the vectorized C matrix with respect
# to the parameters
"derivative" = matrix(numeric(0))[-1,-1]
), # end of C list
# Psi list
Psi = list(
# parameter values of model implied covariance matrix
# numeric matrix of dimension n_ov * n_ov
"values" = matrix(numeric(0))[-1,-1],
# parameter names of model implied covariance matrix
# character matrix of dimension n_ov * n_ov
# same labels display equality constraints
# general: using all-y LISREL notation
# here: using the notation from Gische and Voelkle (2022)
"labels" = matrix(character(0))[-1,-1],
# Partial derivative of the vectorized Psi matrix with respect
# to the parameters
"derivative" = matrix(numeric(0))[-1,-1]
), # end of Psi list
param = list(
# total number of estimated parameters
# normally an integer
# equals the total number of non-NA entries in
# structural_coeff_names and covariance_names
# does NOT take into account other specified equality constraints
"n_par" = integer(0),
# number of distinct and functionally unrelated parameters
# normally an integer
"n_par_unique" = as.integer(0),
# names of distinct and functionally unrelated parameters
# in the order they appear rowwise in the matrices
# structural_coeff_names and covariance_names
# character vector of length(n_par_unique)
"labels_par_unique" = character(0),
# parameter values (estimates) of distinct and functionally unrelated
# parameters
# numeric vector of length(n_par_unique)
"values_par_unique" = numeric(0),
# variance-covariance matrix of estimator of vector
# of distinct and functionally unrelated parameters
# numeric matrix of dimension n_par_unique * n_par_unique
"varcov_par_unique" = matrix(numeric(0))[-1,-1]
) # end of param list
), # end of info_model list
# interventions
info_interventions = list(
# number of interventional variables
# normally an integer
# default: zero
"n_intervention" = as.integer(0),
# names of interventional variables
# character vector of length n_intervention
# default: empty vector (no intervention)
"intervention_names" = character(0),
# number of outcome variables
# normally an integer
# default: n_ov - n_intervention
"n_outcome" = as.integer(0),
# names of outcome variables
# character vector of length(n_outcome)
# default: all non-interventional variables
"outcome_names" = character(0),
# interventional level
# numeric vector of length(n_intervention)
# default: vector of ones
"intervention_levels" = numeric(0),
# parts of the interventional distribution the user is interested in
# character vector
# if "probability" is specified, user specified values of
# lower_bound and upper_bound required
# default: "mean"
# possible values: c("mean", "variance", "density", "probability")
"effect_type" = character(0),
# lower bound of critical range of outcome variables
# numeric vector
"lower_bounds" = numeric(0),
# upper bound of critical range of outcome variables
# numeric vector
"upper_bounds" = numeric(0)
), # end of info_interventions list
# constant matrices for computations
constant_matrices = list(
# selection matrix for entries that are intervened on
# see Definition 1 point 3 in Gische and Voelkle (2022)
# numeric matrix of dimension n_ov x n_intervention
"select_intervention" = matrix(numeric(0))[-1,-1],
# selection matrix for entries that are NOT intervened on
# see Definition 1 point 3 in Gische and Voelkle (2022)
# numeric matrix of dimension n_ov x (n_ov-n_intervention)
"select_non_intervention" = matrix(numeric(0))[-1,-1],
# matrix that replaces entries that are intervened on by zero
# see Definition 1 point 4 in Gische and Voelkle (2022)
# numeric matrix of dimension n_ov x n_ov
"eliminate_intervention" = matrix(numeric(0))[-1,-1],
# selection matrix for outcome variables of interest
# numeric matrix of dimension n_ov x (number of outcome variables
# of interest)
"select_outcome" = matrix(numeric(0))[-1,-1],
# zero-one matrix that maps vech(A) onto vec(A)
# for symmetric A
# numeric matrix of dimension (n_ov^2) x (1/2*n_ov*(n_ov+1))
"duplication_matrix" = matrix(numeric(0))[-1,-1],
# zero-one matrix that eliminates from vec(A)
# the supradiagonal elements of A. In other words,
# it maps vec(A) onto vech(A).
# numeric matrix of dimension (1/2*n_ov*(n_ov+1)) x (n_ov^2)
"elimination_matrix" = matrix(numeric(0))[-1,-1],
# zero-one matrix that transforms vec(A) into vec(A')
# numeric matrix of dimension (n_ov^2) x (n_ov^2)
"commutation_matrix" = matrix(numeric(0))[-1,-1]
), # end of matrices_constant list
# interventional distribution
interventional_distribution = list(
means = list(
# mean vector of the interventional distribution
# see Equation (6a) in Gische and Voelkle (2022)
# numeric vector of length n_ov
"values" = numeric(0),
# jacobian of the mean vector
# numeric matrix of dimension n_ov x n_par_unique
"jacobian" = matrix(numeric(0))[-1,-1],
# asymptotic covariance matrix of the interventional mean vector
# numeric matrix of dimension n_outcomes x n_outcomes
"acov" = matrix(numeric(0))[-1,-1],
# asymptotic standard errors of the interventional means
# numeric vector of length n_ov
"ase" = numeric(0),
# z-values of the interventional means
# numeric vector of length n_ov
"z_values" = numeric(0)
), # end of means
covariance_matrix = list(
# covariance matrix of the interventional distribution
# see Equation (6b) in Gische and Voelkle (2022)
# numeric matrix of dimension n_ov x n_ov
"values" = matrix(numeric(0))[-1,-1],
# jacobian of the VECTORIZED variance-covariance matrix
# numeric matrix of dimension n_ov^2 x n_par_unique
"jacobian" = matrix(numeric(0))[-1,-1],
# asymptotic covariance matrix of the VECTORIZED covariance matrix of the
# interventional distribution
# numeric matrix of dimension n_outcomes^2 x n_outcomes^2
"acov" = matrix(numeric(0))[-1,-1],
# asymptotic standard errors of the interventional VARIANCES
# (i.e., the diagonal entries of the interventional covariance matrix)
# numeric vector of length n n_outcomes
"ase" = matrix(numeric(0))[-1,-1],
# z-values of the the interventional VARIANCES
# (i.e., the diagonal entries of the interventional covariance matrix)
"z_values" = matrix(numeric(0))[-1,-1]
), # end of covariance_matrix
density_function = list(
# probability density function (pdf) of the interventional distribution
# see Equation (9) in Gische and Voelkle (2022)
# numeric matrix of with two columns
# 1 column: grid that captures the range of the outcome variable
# 2 column: values of the pdf
"values" = matrix(numeric(0))[-1,-1],
# jacobian of the density function (pdf) of the interventional
# distribution
# numeric matrix of with two columns
# 1 column: grid that captures the range of the outcome variable
# 2 column: values of the jacobian
"jacobian" = matrix(numeric(0))[-1,-1],
# asymptotic standard errors (ase) of the probability density function
# (pdf) # of the interventional distribution
# numeric matrix of with two columns
# 1 column: grid that captures the range of the outcome variable
# 2 column: values of the ase
"ase" = matrix(numeric(0))[-1,-1],
# z-values of the probability density function (pdf)
# of the interventional distribution
# 1 column: grid that captures the range of the outcome variable
# 2 column: values of the z-values
"z_values" = matrix(numeric(0))[-1,-1]
), # end of density
probabilities = list(
# probabilities of interventional events
# see Equation (10) in Gische and Voelkle (2022)
# numeric vector of length n_outcomes
"values" = numeric(0),
# jacobian of the probability of a UNIVARIATE interventional event
# if n_outcomes > 1, the computation is done separately for each
# outcome variable
# numeric vector of length n_outcomes
"jacobian" = matrix(numeric(0))[-1,-1],
# asymptotic covariance of the probability of a UNIVARIATE
# interventional event
# if n_outcomes > 1, the computation is done separately for each
# outcome variable
# numeric vector of length n_outcomes
"acov" = matrix(numeric(0))[-1,-1],
# asymptotic standard error (ase) of the probability of a UNIVARIATE
# interventional event
# if n_outcomes > 1, the computation is done separately for each
# outcome variable
# numeric vector of length n_outcomes
"ase" = numeric(0),
# asymptotic z-value of the probability of a UNIVARIATE
# interventional event
# if n_outcomes > 1, the computation is done separately for each
# outcome variable
# numeric vector of length n_outcomes
"z_values" = numeric(0)
) # end of probability
), # end of interventional distribution
# control list
control = list(
# verbosity of console output
# a number, 0...no output (default), 1...user messages,
# 2...debugging-relevant messages
"verbose" = verbose
) # end of control list
) # end of internal list
# console output
if( verbose >= 2 ) cat( paste0( " end of function ", fun.name.version, " ",
Sys.time(), "\n" ) )
# return internal list
return( internal_list )
}
## test/development
# source( "c:/Users/martin/Dropbox/68_causalSEM/04_martinhecht/R/
# handle_verbose_argument.R" )
# internal_list <- create_empty_list()
# str( internal_list )
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.