R/RcppExports.R

In goldfish: Statistical Network Models for Dynamic Network Data

# Generated by using Rcpp::compileAttributes() -> do not edit by hand
# Generator token: 10BE3573-1514-4C36-9D1C-5A225CD40393

#' Calculation for estimating an DyNAM-coordination model
#'
#' Output the derivative of log-likelihood, the Fisher information matrix,
#'   the log-Likelihood, and the log-likelihood of each event given input data
#'
#' @noRd
estimate_DyNAM_MM <- function(parameters, dep_event_mat, stat_mat_init, stat_mat_update, stat_mat_update_pointer, presence1_init, presence1_update, presence1_update_pointer, presence2_init, presence2_update, presence2_update_pointer, n_actors_1, n_actors_2, twomode_or_reflexive, impute = TRUE) {
    .Call('_goldfish_estimate_DyNAM_MM', PACKAGE = 'goldfish', parameters, dep_event_mat, stat_mat_init, stat_mat_update, stat_mat_update_pointer, presence1_init, presence1_update, presence1_update_pointer, presence2_init, presence2_update, presence2_update_pointer, n_actors_1, n_actors_2, twomode_or_reflexive, impute)
}

#' Calculation for estimating an DyNAM choice model
#' @noRd
estimate_DyNAM_choice <- function(parameters, dep_event_mat, stat_mat_init, stat_mat_update, stat_mat_update_pointer, presence2_init, presence2_update, presence2_update_pointer, n_actors_1, n_actors_2, twomode_or_reflexive, impute) {
    .Call('_goldfish_estimate_DyNAM_choice', PACKAGE = 'goldfish', parameters, dep_event_mat, stat_mat_init, stat_mat_update, stat_mat_update_pointer, presence2_init, presence2_update, presence2_update_pointer, n_actors_1, n_actors_2, twomode_or_reflexive, impute)
}

#' Calculation for estimating an DyNAM-rate model
#' @noRd
estimate_DyNAM_rate <- function(parameters, dep_event_mat, timespan, is_dependent, stat_mat_init, stat_mat_update, stat_mat_update_pointer, stat_mat_rightcensored_update, stat_mat_rightcensored_update_pointer, presence1_init, presence1_update, presence1_update_pointer, presence2_init, presence2_update, presence2_update_pointer, n_actors_1, n_actors_2, twomode_or_reflexive, impute = TRUE) {
    .Call('_goldfish_estimate_DyNAM_rate', PACKAGE = 'goldfish', parameters, dep_event_mat, timespan, is_dependent, stat_mat_init, stat_mat_update, stat_mat_update_pointer, stat_mat_rightcensored_update, stat_mat_rightcensored_update_pointer, presence1_init, presence1_update, presence1_update_pointer, presence2_init, presence2_update, presence2_update_pointer, n_actors_1, n_actors_2, twomode_or_reflexive, impute)
}

#' Calculation for estimating an DyNAM-rate-ordered model
#' @noRd
estimate_DyNAM_rate_ordered <- function(parameters, dep_event_mat, stat_mat_init, stat_mat_update, stat_mat_update_pointer, presence1_init, presence1_update, presence1_update_pointer, presence2_init, presence2_update, presence2_update_pointer, n_actors_1, n_actors_2, twomode_or_reflexive, impute = TRUE) {
    .Call('_goldfish_estimate_DyNAM_rate_ordered', PACKAGE = 'goldfish', parameters, dep_event_mat, stat_mat_init, stat_mat_update, stat_mat_update_pointer, presence1_init, presence1_update, presence1_update_pointer, presence2_init, presence2_update, presence2_update_pointer, n_actors_1, n_actors_2, twomode_or_reflexive, impute)
}

#' Calculation for estimating an REM-choice model
#'
#' Given input data, it output the derivative of the loglikelihood,
#'    the Fisher information matrix, the logLikelihood, 
#'    and the loglikelihood of each event
#'
#' @param parameters An n_parameters by 1 matrix, which is the input parameter
#' @param dep_event_mat An 2 by n_events matrix, the (1,n) entry is 
#'    the sender of n-th event, and the (2,n) entry is the 
#'    receiver of the n-th event.
#' @param timespan An n_events by 1 matrix.
#'    The i-th element is the waiting time of the i-th event.
#' @param is_dependent An n_events by 1 matrix with boolean values.
#'    If the i-th event is dependent, then the i-th entry of
#'    is_dependent is TRUE, otherwise it's FALSE.
#' @param stat_mat_init An n_actor1*n_actor2 by n_parameters matrix.
#'    It is the initialization of the statistics matrix.
#'    The initial value of k-th effect for the actor1-actor2 pair (i,j)
#'    is recorded in the ((i-1)*n_actor2 + j,k) entry of stat_mat_init.
#' @param stat_mat_update An matrix with four rows,
#'    which records the updates of the statistics matrix through all
#'    dependent events.
#'    The following is an example.
#'     \tabular{rrrrr}{
#'       0 \tab 2 \tab 0 \tab 9 \tab 9\cr
#'       2 \tab 1 \tab 5 \tab 4 \tab 3\cr
#'       3 \tab 5 \tab 5 \tab  0 \tab 5\cr
#'       1.2 \tab 3.5 \tab 2.5 \tab 9.23 \tab 2.8\cr
#'     }
#'     Each column represents an update.
#'     For example the first column means replacing the value of
#'     the (3+1)-th effect for the actor1-actor2 pair(0+1,2+1) by 1.2.
#'     The +1 is due to the difference between the numberings in R and C.
#' @param stat_mat_update_pointer An n_events by 1 matrix that record
#'     which update belongs to which dependent event.
#'     Suppose that the first three elements are (10,11,15).
#'     Then the first 10 colums of stat_mat_update is the update for 
#'     the first event, the 11th column is  the update for the second event,
#'     and the 12th to 15th columns are the updates for the third event.
#' @param stat_mat_rightcensored_update An matrix with four rows,
#'     which record the updates of the statistics matrix through all 
#'     right-censored events.
#'     the structure is similar to stat_mat_update.
#' @param stat_mat_update_pointer An n_events by 1 matrix that record 
#'    which update belongs to which rightcensored event.
#'    The structure is similar to stat_mat_update_pointer.
#' @param presence1_init An n_actor1 by 1 matrix, which records 
#'    the initial presence of each actor1.
#'    If the i-th actor1 is not present in the
#'    beginning then the i-th entry of presence1_init is 0, otherwise it's 1.
#' @param presence1_update An matrix with two rows, which record the 
#'    updates of the presence of actor1 through all events.
#'    The following is an example.
#'     \tabular{rrrrr}{
#'       0 \tab 3 \tab 4 \tab 9 \tab 20\cr
#'       1 \tab 0 \tab 1 \tab 0 \tab 0\cr
#'     }
#'    Each column represents an update. For example the first column means
#'    the 0+1-th actor1 becomes present.
#'    And the second column means the
#'    the 3+1 th actor1 becomes absent. 
#'    The +1 is due to the difference between the numberings in R and C.
#' @param presence2_update_pointer An n_events by 1 matrix that record 
#'    which update belongs to which (dependent+ rightcensored) event.
#'    The structure is similar to stat_mat_update_pointer.
#' @param presence2_init An n_actors2 by 1 matrix, which records the
#'    initial presence of each actor2.
#'    If the i-th actor2 is not present in the beginning then the i-th entry
#'    of presence2_init is 0, otherwise it's 1.
#' @param presence2_update An matrix with two rows, which record the updates
#'    of the presence of actor2 through all events.
#'    The following is an example.
#'     \tabular{rrrrr}{
#'       0 \tab 3 \tab 4 \tab 9 \tab 20\cr
#'       1 \tab 0 \tab 1 \tab 0 \tab 0\cr
#'     }
#'    Each column represents an update. For example the first column means
#'    the 0+1-th actor2 becomes present.
#'    And the second column means the 3+1 th actor2 becomes absent.
#'    The +1 is due to the difference between the numberings in R and C.
#' @param presence2_update_pointer An n_events by 1 matrix that record
#'    which update belongs to which (dependent+ rightcensored) event.
#'    The structure is similar to stat_mat_update_pointer.
#' @param n_actors_1 An integer which is the number of actor1
#' @param n_actors_2 An integer which is the number of actor2
#' @param twomode_or_reflexive An boolean variable. If it's true,
#'    then the model is two-model or we consider the reflexive effect
#'    (that's the value in the diagonal entries of the statistics matrix).
#' @param impute An boolean variable. If it true, the function does
#'    the imputation for missing values.
#'
#' @return Return a list with elements as follows.
#' \describe{
#'   \item{derivative}{An 1 by n_parameters matrix, which is the derivative of
#'    loglikelihood given the input parameter and data.}
#'   \item{fisher}{An n_parameters by n_parameters matrix, which is the fisher
#'    information matrix given the input parameter and data.}
#'   \item{logLikelihood}{An scalar, which is the loglikelihood given
#'   the input parameter and data.}
#'   \item{intervalLogL}{An n_events by 1 matrix,
#'         of which the i-th entry is the loglikelihood of the i-th event
#'         given the input parameter and data.}
#' }
#' @noRd
estimate_REM <- function(parameters, dep_event_mat, timespan, is_dependent, stat_mat_init, stat_mat_update, stat_mat_update_pointer, stat_mat_rightcensored_update, stat_mat_rightcensored_update_pointer, presence1_init, presence1_update, presence1_update_pointer, presence2_init, presence2_update, presence2_update_pointer, n_actors_1, n_actors_2, twomode_or_reflexive, impute) {
    .Call('_goldfish_estimate_REM', PACKAGE = 'goldfish', parameters, dep_event_mat, timespan, is_dependent, stat_mat_init, stat_mat_update, stat_mat_update_pointer, stat_mat_rightcensored_update, stat_mat_rightcensored_update_pointer, presence1_init, presence1_update, presence1_update_pointer, presence2_init, presence2_update, presence2_update_pointer, n_actors_1, n_actors_2, twomode_or_reflexive, impute)
}

#' Calculation for estimating an REM-choice-ordered model
#' @noRd
estimate_REM_ordered <- function(parameters, dep_event_mat, stat_mat_init, stat_mat_update, stat_mat_update_pointer, presence1_init, presence1_update, presence1_update_pointer, presence2_init, presence2_update, presence2_update_pointer, n_actors_1, n_actors_2, twomode_or_reflexive, impute = TRUE) {
    .Call('_goldfish_estimate_REM_ordered', PACKAGE = 'goldfish', parameters, dep_event_mat, stat_mat_init, stat_mat_update, stat_mat_update_pointer, presence1_init, presence1_update, presence1_update_pointer, presence2_init, presence2_update, presence2_update_pointer, n_actors_1, n_actors_2, twomode_or_reflexive, impute)
}

#' Estimate a DyNAM-coordination model with gathered data
#'
#' Given the gathered and distilled data, it outputs the derivative of
#'     the loglikelihood, the Fisher information matrix, the logLikelihood,
#'     and the loglikelihood of each event  for DyNAM-coordination models.
#'
#' @param parameters An n_effects by 1 matrix, which is the input parameter
#' @param stat_all_events An matrix with n_effects columns.
#'     Each row represent the values of all effects of
#'     a sender-receiver pair in an event.
#'     For example, for a model with 2 effects, stat_all_events 
#'     might looks like this.
#'     \tabular{rr}{
#'       3.2 \tab 1.9\cr
#'       3.2 \tab 4.5\cr
#'       1.2 \tab 5.2\cr
#'       4.3 \tab 3.1\cr
#'       2.4 \tab 4.7\cr
#'       9.2 \tab 5.6\cr
#'       2.9 \tab 8.9\cr
#'       ... \tab ...\cr
#'     }
#'     The first row means in a event, the values of the two effects of
#'     a candidate pair is (3.2,1.9).
#' @param n_candidates An n_events by 1 matrix.
#'     It record how many candidate sender-receiver pairs are in each event.
#'     It indicates which row in stat_all_events belongs to which event.
#'     For example if the first two element of n_candidates are (2, 4),
#'     then the first 2 rows of stat_all_events correspond to
#'     the candidate pairs in the first event, which is
#'     \tabular{rr}{
#'       3.2 \tab 1.9\cr
#'       3.2 \tab 4.5\cr
#'     }
#'     And the 3rd-6th rows correspond to the candidate pairs in
#'     the second event, which is
#'     \tabular{rr}{
#'       1.2 \tab 5.2\cr
#'       4.3 \tab 3.1\cr
#'       2.4 \tab 4.7\cr
#'       9.2 \tab 5.6\cr
#'     }
#' @param n_candidates1 An n_events by 1 matrix, which is only used
#'     for estimating the DyNAM-coordination model.
#'     It record how many candidate sender are in each event.
#'     And we have n_candidates1 * n_candidates2 = n_candidates.
#' @param n_candidates2 An n_events by 1 matrix, which is only used
#'     for estimating the DyNAM-coordination model.
#'     It record how many candidate receiver are in each event.
#'     And we have n_candidates1 * n_candidates2 = n_candidates.
#' @param selected An n_events by 1 matrix.
#'     It records the position of the selected candidate sender-receiver pair
#'     in each event.
#'     For example  if the first two element of n_candidates are 2, 4,
#'     and the first two elements of selected is (0,2).
#'     Then the value of the effects of the pair selected in the first event is
#'     (3.2,1.9), which is the 1st(=0+1) row of
#'     \tabular{rr}{
#'       3.2 \tab 1.9\cr
#'       3.2 \tab 4.5\cr
#'     }
#'     And the value of the effects of the pair selected in the second event is
#'     (2.4, 4.7), which is the 3rd(=2+1) row of
#'     \tabular{rr}{
#'       1.2 \tab 5.2\cr
#'       4.3 \tab 3.1\cr
#'       2.4 \tab 4.7\cr
#'       9.2 \tab 5.6\cr
#'     }
#' @param selected_actor1 An n_events by 1 matrix.
#'     It records the index of the selected candidate sender among
#'     all candidate sender in each event.
#' @param selected_actor2 An n_events by 1 matrix.
#'     It records the index of the selected candidate receiver among
#'     all candidate receiver in each event.
#' @noRd
compute_coordination_selection <- function(parameters, stat_all_events, n_candidates, n_candidates1, n_candidates2, selected, selected_actor1, selected_actor2, twomode_or_reflexive) {
    .Call('_goldfish_compute_coordination_selection', PACKAGE = 'goldfish', parameters, stat_all_events, n_candidates, n_candidates1, n_candidates2, selected, selected_actor1, selected_actor2, twomode_or_reflexive)
}

#' Estimate a multinomial selection model with gathered data
#'
#' Given the gathered and distilled data, it outputs the derivative of
#'   the log-likelihood, the Fisher information matrix, the log-Likelihood,
#'   and the log-likelihood of each event for models with
#'   multinomial selection processes, e.g. DyNAM-rate-ordered, DyNAM-choice,
#'   and REM-choice models.
#' @noRd
compute_multinomial_selection <- function(parameters, stat_all_events, n_candidates, selected) {
    .Call('_goldfish_compute_multinomial_selection', PACKAGE = 'goldfish', parameters, stat_all_events, n_candidates, selected)
}

#' Estimate a poisson selection model with gathered data
#'
#' Given the gathered and distilled data,
#' it outputs the derivative of the log-likelihood,
#' the Fisher information matrix, the log-likelihood,
#' and the log-likelihood of each event 
#' for models with poisson selection processes,
#' e.g., DyNAM-rate and REM-choice models.
#' @noRd
compute_poisson_selection <- function(parameters, stat_all_events, n_candidates, selected, timespan, is_dependent) {
    .Call('_goldfish_compute_poisson_selection', PACKAGE = 'goldfish', parameters, stat_all_events, n_candidates, selected, timespan, is_dependent)
}

#' a function to extract the update of composition change
#'   from an events and transform the data into a matrix
#'   and a vector.
#'
#' @param event an event objects with the information on composition change
#' @param reference_event_time a vector of time stamps that separate updates
#'   different time spans.
#' @return A list with two element: changeMat and change_idx.
#' For the structure of these two object see, e.g.,
#'   the documentation of estimate_DyNAM_choice,
#'   in which they are used for stat_mat_update, and stat_mat_update_pointer.
#' @noRd
C_convert_composition_change <- function(event, reference_event_time) {
    .Call('_goldfish_C_convert_composition_change', PACKAGE = 'goldfish', event, reference_event_time)
}

convert_composition_change <- function(event, reference_event) {
    .Call('_goldfish_convert_composition_change', PACKAGE = 'goldfish', event, reference_event)
}

#' a function to transform a list of updates into a matrix
#' and a vector.
#' @noRd
convert_change <- function(changeList) {
    .Call('_goldfish_convert_change', PACKAGE = 'goldfish', changeList)
}

#' Gathering data for sender receiver model
#'
#' Gathering data for model that for choosing an receiver given a sender,
#' i.e. DyNAM-choice model.
#' Only the useful information are recorded: Consider a actor1-actor2 pair,
#' if the actor1 is not the sender of the current event
#' or the actor2 is not present, then the information is not collected.
#' @param verbose An boolean variable. It it's true, the function print
#' the progress to the screen.
#' @return Return a list with elements as follows. The meaning of the argument
#' can be found in corresponding computation codes,
#' e.g. compute_coordination_selection.cpp.
#' @noRd
gather_receiver_model <- function(dep_event_mat, stat_mat_init, stat_mat_update, stat_mat_update_pointer, presence2_init, presence2_update, presence2_update_pointer, n_actor1, n_actor2, twomode_or_reflexive, verbose = FALSE, impute = TRUE) {
    .Call('_goldfish_gather_receiver_model', PACKAGE = 'goldfish', dep_event_mat, stat_mat_init, stat_mat_update, stat_mat_update_pointer, presence2_init, presence2_update, presence2_update_pointer, n_actor1, n_actor2, twomode_or_reflexive, verbose, impute)
}

#' Gathering data for sender receiver model
#'
#' Gathering data for models for choosing an sender,
#'     i.e. DyNAM-rate and DyNAM-rate-ordered models.
#'     Only the useful information are recorded,
#'     e.g. if a actor1 is not present then its information is not collected.
#' @param verbose An boolean variable. It it's true,
#'     the function prints the progress to the screen.
#' @return Return a list with elements as follows.
#'     The meaning of the argument can be found in corresponding
#'     computation codes, e.g. compute_poisson_selection.cpp.
#' @noRd
gather_sender_model <- function(dep_event_mat, is_dependent, stat_mat_init, stat_mat_update, stat_mat_update_pointer, stat_mat_rightcensored_update, stat_mat_rightcensored_update_pointer, presence1_init, presence1_update, presence1_update_pointer, presence2_init, presence2_update, presence2_update_pointer, n_actors_1, n_actors_2, twomode_or_reflexive, verbose, impute = TRUE) {
    .Call('_goldfish_gather_sender_model', PACKAGE = 'goldfish', dep_event_mat, is_dependent, stat_mat_init, stat_mat_update, stat_mat_update_pointer, stat_mat_rightcensored_update, stat_mat_rightcensored_update_pointer, presence1_init, presence1_update, presence1_update_pointer, presence2_init, presence2_update, presence2_update_pointer, n_actors_1, n_actors_2, twomode_or_reflexive, verbose, impute)
}

#' Gathering data for sender receiver model
#'
#' Gathering data for model that considers all present sender-receiver pairs,
#'   i.e. REM, REM-ordered, and DyNAM-coordination models.
#'   Only the useful information are recorded, 
#'   e.g. if a actor1-actor2 pair is not present in an event,
#'   then its information is  not gathered by this function.
#' @param verbose An boolean variable.
#'   It it's true, the function print the progress to the screen.
#' @return Return a list with elements as follows.
#' The meaning of the argument can be found in corresponding computation codes,
#' e.g. compute_coordination_selection.cpp.
#' @noRd
gather_sender_receiver_model <- function(dep_event_mat, is_dependent, stat_mat_init, stat_mat_update, stat_mat_update_pointer, stat_mat_rightcensored_update, stat_mat_rightcensored_update_pointer, presence1_init, presence1_update, presence1_update_pointer, presence2_init, presence2_update, presence2_update_pointer, n_actors_1, n_actors_2, twomode_or_reflexive, verbose, impute) {
    .Call('_goldfish_gather_sender_receiver_model', PACKAGE = 'goldfish', dep_event_mat, is_dependent, stat_mat_init, stat_mat_update, stat_mat_update_pointer, stat_mat_rightcensored_update, stat_mat_rightcensored_update_pointer, presence1_init, presence1_update, presence1_update_pointer, presence2_init, presence2_update, presence2_update_pointer, n_actors_1, n_actors_2, twomode_or_reflexive, verbose, impute)
}