PcalgSearch: R6 Interface to pcalg Search Algorithms

In causalDisco: Tools for Causal Discovery on Observational Data

R6 Interface to pcalg Search Algorithms

Description

A wrapper that lets you drive pcalg algorithms within the causalDisco framework. For arguments to the test, score, and algorithm, see the pcalg documentation, which we link to in the respective sections below.

Public fields

data

A data.frame holding the data set currently attached to the search object. Can be set with set_data().

score

A function that will be used to build the score, when data is set. Can be set with $set_score(). Recognized values are:

• "sem_bic" - BIC score for Gaussian observed data. See pcalg::GaussL0penObsScore.

• "sem_bic_int" - BIC score for Gaussian data from jointly interventional and observational Gaussian data. See pcalg::GaussL0penIntScore.

test

A function that will be used to test independence. Can be set with $set_test(). Recognized values are:

• "fisher_z" - Fisher Z test for Gaussian data. See pcalg::gaussCItest().

• "fisher_z_twd" - Fisher Z test for Gaussian data with test-wise deletion. See micd::gaussCItwd().

• "fisher_z_mi" - Fisher Z test for Gaussian data with multiple imputation. See micd::gaussCItestMI().

• "g_square" - G square test for discrete data. See pcalg::binCItest() and pcalg::disCItest().

• "g_square_twd" - G square test for discrete data with test-wise deletion. See micd::disCItwd().

• "g_square_mi" - G square test for discrete data with multiple imputation. See micd::disMItest().

• "conditional_gaussian" - Test for conditional independence in mixed data. See micd::mixCItest().

• "conditional_gaussian_twd" - Test for conditional independence in mixed data with test-wise deletion. See micd::mixCItwd().

• "conditional_gaussian_mi" - Test for conditional independence in mixed data with multiple imputation. See micd::mixMItest().

alg

A function that will be used to run the search algorithm. Can be set with $set_alg(). Recognized values are:

• "fci" - FCI algorithm. See fci() and the underlying pcalg::fci().

• "ges" - GES algorithm. See ges() and the underlying pcalg::ges().

• "pc" - PC algorithm. See pc() and the underlying pcalg::pc().

params

A list of parameters for the test and algorithm. Can be set with $set_params(). The parameters are passed to the test and algorithm functions.

suff_stat

Sufficient statistic. The format and contents of the sufficient statistic depends on which test is being used.

continuous

Logical; whether the sufficient statistic is for a continuous test. If both continuous and discrete are TRUE, the sufficient statistic is build for a mixed test.

discrete

Logical; whether the sufficient statistic is for a discrete test. If both continuous and discrete are TRUE, the sufficient statistic is build for a mixed test.

knowledge

A list of fixed constraints for the search algorithm. Note, that pcalg only works with symmetric knowledge. Thus, the only allowed types of knowledge is forbidden edges in both directions.

adapt_df

Logical; whether to adapt the degrees of freedom for discrete tests.

Methods

Public methods

• PcalgSearch$new()

• PcalgSearch$set_params()

• PcalgSearch$set_data()

• PcalgSearch$set_suff_stat()

• PcalgSearch$set_test()

• PcalgSearch$set_score()

• PcalgSearch$set_alg()

• PcalgSearch$set_knowledge()

• PcalgSearch$run_search()

• PcalgSearch$clone()

---

Method new()

Constructor for the PcalgSearch class.

Usage

PcalgSearch$new()

---

Method set_params()

Sets the parameters for the test and algorithm.

Usage

PcalgSearch$set_params(params)

Arguments

params

A list of parameters to set.

---

Method set_data()

Sets the data for the search algorithm.

Usage

PcalgSearch$set_data(data, set_suff_stat = TRUE)

Arguments

data

A data.frame or a matrix containing the data.

set_suff_stat

Logical; whether to set the sufficient statistic. for the data.

---

Method set_suff_stat()

Sets the sufficient statistic for the data.

Usage

PcalgSearch$set_suff_stat()

---

Method set_test()

Sets the test for the search algorithm.

Usage

PcalgSearch$set_test(method, alpha = 0.05, suff_stat_fun = NULL, args = NULL)

Arguments

method

A string specifying the type of test to use.

Can also be a user-defined function with signature ⁠function(x, y, conditioning_set, suff_stat)⁠, where x and y are the variables being tested for independence, conditioning_set is the conditioning set, and suff_stat is the sufficient statistic for the test. If a user-defined function is provided, then suff_stat_fun must also be provided, which is a function that should take the data as input and returns a sufficient statistic for the test. Optionally, the signature of the user-defined test function can also include an args parameter, which is a list of additional arguments to pass to the test function. If args is provided, then the test function should have the signature ⁠function(x, y, conditioning_set, suff_stat, args)⁠, and the args parameter will be passed to the test function.

EXPERIMENTAL: user-defined tests syntax are subject to change.

alpha

Significance level for the test.

suff_stat_fun

A function that takes the data as input and returns a sufficient statistic for the test. Only needed if method is a user-defined function.

args

A list of additional arguments to pass to the test. Only needed if method is a user-defined function with an args parameter in its signature.

---

Method set_score()

Sets the score for the search algorithm.

Usage

PcalgSearch$set_score(method, params = list())

Arguments

method

A string specifying the type of score to use.

params

A list of parameters to pass to the score function.

---

Method set_alg()

Sets the algorithm for the search.

Usage

PcalgSearch$set_alg(method)

Arguments

method

A string specifying the type of algorithm to use.

---

Method set_knowledge()

Sets the knowledge for the search algorithm. Due to the nature of pcalg, we cannot set knowledge before we run it on data. So we set the function that will be used to build the fixed constraints, but it can first be done when data is provided.

Usage

PcalgSearch$set_knowledge(knowledge_obj, directed_as_undirected = FALSE)

Arguments

knowledge_obj

A Knowledge object that contains the fixed constraints.

directed_as_undirected

Logical; whether to treat directed edges as undirected.

---

Method run_search()

Runs the search algorithm on the data.

Usage

PcalgSearch$run_search(data = NULL, set_suff_stat = TRUE)

Arguments

data

A data.frame or a matrix containing the data.

set_suff_stat

Logical; whether to set the sufficient statistic

---

Method clone()

The objects of this class are cloneable with this method.

Usage

PcalgSearch$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Examples

### pcalg_search R6 class examples ###

# Generally, we do not recommend using the R6 classes directly, but rather
# use the disco() or any method function, for example pc(), instead.

# Load data
data(num_data)

# Recommended:
my_pc <- pc(engine = "pcalg", test = "fisher_z")
my_pc(num_data)

# or
disco(data = num_data, method = my_pc)

# Example with detailed settings:
my_pc2 <- pc(
  engine = "pcalg",
  test = "fisher_z",
  alpha = 0.01,
  m.max = 4,
  skel.method = "original"
)

disco(data = num_data, method = my_pc2)

# With knowledge

kn <- knowledge(
  num_data,
  X1 %!-->% X2,
  X2 %!-->% X1
)

disco(data = num_data, method = my_pc2, knowledge = kn)

# Using R6 class:
s <- PcalgSearch$new()

s$set_test(method = "fisher_z", alpha = 0.05)
s$set_data(tpc_example)
s$set_alg("pc")

g <- s$run_search()

print(g)

causalDisco documentation built on April 13, 2026, 5:06 p.m.