R6 Interface to Tetrad Search Algorithms

High-level wrapper around the Java-based Tetrad causal-discovery library. The class lets you choose independence tests, scores, and search algorithms from Tetrad, run them on an R data set, and retrieve the resulting graph or statistics.

Public fields

data

Java object that stores the (possibly converted) data set used by Tetrad.

rdata

Original R data.frame supplied by the user.

score

Java object holding the scoring function selected with set_score(). Supply one of the method strings for set_score(). Recognised values are:

Continuous - Gaussian

"ebic" - Extended BIC score.
"gic" - Generalized Information Criterion (GIC) score.
"poisson_prior" - Poisson prior score.'
"rank_bic" - Rank-based BIC score.
"sem_bic" - SEM BIC score.
"zhang_shen_bound" - Zhang and Shen bound score.

Discrete - categorical

"bdeu" - Bayes Dirichlet Equivalent score with uniform priors.
"discrete_bic" - BIC score for discrete data.

Mixed Discrete/Gaussian

"basis_function_bic" - BIC score for basis-function models. This is a generalization of the Degenerate Gaussian score.
"basis_function_blocks_bic" - BIC score for mixed data using basis-function models.
"basis_function_sem_bic" - SEM BIC score for basis-function models.
"conditional_gaussian" - Conditional Gaussian BIC score.
"degenerate_gaussian" - Degenerate Gaussian BIC score.
"mag_degenerate_gaussian_bic" - MAG Degenerate Gaussian BIC Score.

test

Java object holding the independence test selected with set_test(). Supply one of the method strings for set_test(). Recognised values are:

Continuous - Gaussian

"fisher_z" - Fisher $Z$ (partial correlation) test.
"poisson_prior" - Poisson prior test.
"rank_independence" - Rank-based independence test.
"sem_bic" - SEM BIC test.

Discrete - categorical

"chi_square" - chi-squared test
"g_square" - likelihood-ratio $G^2$ test.
"probabilistic" - Uses BCInference by Cooper and Bui to calculate probabilistic conditional independence judgments.

General

"gin" - Generalized Independence Noise test.
"kci" - Kernel Conditional Independence Test (KCI) by Kun Zhang.
"rcit" - Randomized Conditional Independence Test (RCIT).

Mixed Discrete/Gaussian

"basis_function_blocks" - Basis-function blocks test.
"basis_function_lrt" - basis-function likelihood-ratio.
"conditional_gaussian" - Conditional Gaussian test as a likelihood ratio test.
"degenerate_gaussian" - Degenerate Gaussian test as a likelihood ratio test.

alg

Java object representing the search algorithm. Supply one of the method strings for set_alg(). Recognised values are:

Constraint-based

"fci" - FCI algorithm. See fci().
"pc" - Peter-Clark (PC) algorithm. See pc().
"rfci" - Restricted FCI algorithm. See pcalg::rfci().

Hybrid

"boss_fci" - BOSS-FCI algorithm. See boss_fci().
"gfci" - GFCI algorithm. See gfci().
"grasp_fci" - GRaSP-FCI algorithm. See grasp_fci().
"sp_fci" - Sparsest Permutation using FCI. See sp_fci().

Score-based

"boss" - BOSS algorithm. See boss().
"ges" ("fges") - (Fast) Greedy Equivalence Search (GES) algorithm. See ges().
"grasp" - GRaSP (Greedy Relations of Sparsest Permutation) algorithm. See grasp().

mc_test

Java independence-test object used by the Markov checker.

java

Java object returned by the search (typically a graph).

result

Convenience alias for java; may store additional metadata depending on the search type.

knowledge

Java Knowledge object carrying background constraints (required/forbidden edges).

params

Java Parameters object holding algorithm settings.

bootstrap_graphs

Java List of graphs produced by bootstrap resampling, if that feature was requested.

mc_ind_results

Java List with Markov-checker test results.

Methods

Method `new()`

Initializes the TetradSearch object, creating new Java objects for knowledge and params.

Usage

TetradSearch$new()

Method `set_test()`

Sets the independence test to use in Tetrad.

Usage

TetradSearch$set_test(method, ..., use_for_mc = FALSE)

Arguments

method

(character) Name of the test method (e.g., "chi_square", "fisher_z").

"basis_function_blocks" - Basis-function blocks test
"basis_function_lrt" - basis-function likelihood-ratio
"chi_square" - chi-squared test
"conditional_gaussian" - Mixed discrete/continuous test
"degenerate_gaussian" - Degenerate Gaussian test as a likelihood ratio test
"fisher_z" - Fisher $Z$ (partial correlation) test
"gin" - Generalized Independence Noise test
"kci" - Kernel Conditional Independence Test (KCI) by Kun Zhang
"poisson_prior" - Poisson prior test
"probabilistic" - Uses BCInference by Cooper and Bui to calculate probabilistic conditional independence judgments.
"rcit" - Randomized Conditional Independence Test (RCIT)
"rank_independence" - Rank-based independence test
"sem_bic" - SEM BIC test

...

Additional arguments passed to the private test-setting methods. For the following tests, the following parameters are available:

"basis_function_blocks" - Basis-function blocks test.
- alpha = 0.05 - Significance level for the independence test,
- basis_type = "polynomial" - The type of basis to use. Supported types are "polynomial", "legendre", "hermite", and "chebyshev",
- truncation_limit = 3 - Basis functions 1 through this number will be used. The Degenerate Gaussian category indicator variables for mixed data are also used.
"basis_function_lrt" - basis-function likelihood-ratio
- truncation_limit = 3 - Basis functions 1 through this number will be used. The Degenerate Gaussian category indicator variables for mixed data are also used,
- alpha = 0.05 - Significance level for the likelihood-ratio test,
- singularity_lambda = 0.0 - Small number >= 0: Add lambda to the diagonal, < 0 Pseudoinverse,
- do_one_equation_only = FALSE - If TRUE, only one equation should be used when expanding the basis.
"chi_square" - chi-squared test
- min_count = 1 - Minimum count for the chi-squared test per cell. Increasing this can improve accuracy of the test estimates,
- alpha = 0.05 - Significance level for the independence test,
- cell_table_type = "ad" - The type of cell table to use for optimization. Available types are: "ad" - AD tree, "count" - Count sample.
"conditional_gaussian" - Mixed discrete/continuous test
- alpha = 0.05 - Significance level for the independence test,
- discretize = TRUE - If TRUE for the conditional Gaussian likelihood, when scoring X –> D where X is continuous and D discrete, one should to simply discretize X for just those cases. If FALSE, the integration will be exact,
- num_categories_to_discretize = 3 - In case the exact algorithm is not used for discrete children and continuous parents is not used, this parameter gives the number of categories to use for this second (discretized) backup copy of the continuous variables,
- min_sample_size_per_cell = 4 - Minimum sample size per cell for the independence test.
"degenerate_gaussian" - Degenerate Gaussian likelihood ratio test
- alpha = 0.05 - Significance level for the independence test,
- singularity_lambda = 0.0 - Small number >= 0: Add lambda to the diagonal, < 0 Pseudoinverse.
"fisher_z" - Fisher $Z$ (partial correlation) test
- alpha = 0.05 - Significance level for the independence test,
- singularity_lambda = 0.0 - Small number >= 0: Add lambda to the diagonal, < 0 Pseudoinverse.
"gin" - Generalized Independence Noise test.
- alpha = 0.05 - Significance level for the independence test,
- gin_backend = "dcor" - Unconditional test for residual independence. Available types are "dcor" - Distance correlation (for non-linear) and "pearson" - Pearson correlation (for linear),
- num_permutations = 200 - Number of permutations used for "dcor" backend. If "pearson" backend is used, this parameter is ignored.
- gin_ridge = 1e-8 - Ridge parameter used when computing residuals. A small number >= 0.
- seed = -1 - Random seed for the independence test. If -1, no seed is set.
"kci" - Kernel Conditional Independence Test (KCI) by Kun Zhang
- alpha = 0.05 - Significance level for the independence test,
- approximate = TRUE - If TRUE, use the approximate Gamma approximation algorithm. If FALSE, use the exact,
- scaling_factor = 1 - For Gaussian kernel: The scaling factor * Silverman bandwidth.
- num_bootstraps = 1000 - Number of bootstrap samples to use for the KCI test. Only used if approximate = FALSE.
- threshold = 1e-3 - Threshold for the KCI test. Threshold to determine how many eigenvalues to use – the lower the more (0 to 1).
- kernel_type = "gaussian" - The type of kernel to use. Available types are "gaussian", "linear", or "polynomial".
- polyd = 5 - The degree of the polynomial kernel, if used.
- polyc = 1 - The constant of the polynomial kernel, if used.
"poisson_prior" - Poisson prior test
- poisson_lambda = 1 - Lambda parameter for the Poisson distribution (> 0),
- singularity_lambda = 0.0 - Small number >= 0: Add lambda to the diagonal, < 0 Pseudoinverse.
"probabilistic" - Uses BCInference by Cooper and Bui to calculate probabilistic conditional independence judgments.
- threshold = FALSE - Set to TRUE if using the cutoff threshold for the independence test,
- cutoff = 0.5 - Cutoff for the independence test,
- prior_ess = 10 - Prior equivalent sample size for the independence test. This number is added to the sample size for each conditional probability table in the model and is divided equally among the cells in the table.
"rcit" - Randomized Conditional Independence Test (RCIT).
- alpha = 0.05 - Significance level for the independence test,
- rcit_approx = "lpb4" - Null approximation method. Recognized values are: "lpb4" - Lindsay-Pilla-Basak method with 4 support points, "hbe" - Hall-Buckley-Eagleson method, "gamma" - Gamma (Satterthwaite-Welch), "chi_square" - Chi-square (normalized), "permutation" - Permutation-based (computationally intensive),
- rcit_ridge = 1e-3 - Ridge parameter used when computing residuals. A small number >= 0,
- num_feat = 10 - Number of random features to use for the regression of X and Y on Z. Values between 5 and 20 often suffice.
- num_fourier_feat_xy = 5 - Number of random Fourier features to use for the tested variables X and Y. Small values often suffice (e.g., 3 to 10),
- num_fourier_feat_z = 100 - Number of random Fourier features to use for the conditioning set Z. Values between 50 and 300 often suffice,
- center_features = TRUE - If TRUE, center the random features to have mean zero. Recommended for better numerical stability,
- use_rcit = TRUE - If TRUE, use RCIT; if FALSE, use RCoT (Randomized Conditional Correlation Test),
- num_permutations = 500 - Number of permutations used for the independence test when rcit_approx = "permutation" is selected.
- seed = -1 - Random seed for the independence test. If -1, no seed is set.
"rank_independence" - Rank-based independence test.
- alpha = 0.05 - Significance level for the independence test.
"sem_bic" - SEM BIC test.
- penalty_discount = 2 - Penalty discount factor used in BIC = 2L - ck log N, where c is the penalty. Higher c yield sparser graphs,
- structure_prior = 0 - The default number of parents for any conditional probability table. Higher weight is accorded to tables with about that number of parents. The prior structure weights are distributed according to a binomial distribution,
- singularity_lambda = 0.0 - Small number >= 0: Add lambda to the diagonal, < 0 Pseudoinverse.

use_for_mc

(logical) If TRUE, sets this test for the Markov checker mc_test.

Returns

Invisibly returns self, for chaining.

Method `set_score()`

Sets the scoring function to use in Tetrad.

Usage

TetradSearch$set_score(method, ...)

Arguments

method

(character) Name of the score (e.g., "sem_bic", "ebic", "bdeu").

"bdeu" - Bayes Dirichlet Equivalent score with uniform priors.
"basis_function_bic" - BIC score for basis-function models. This is a generalization of the Degenerate Gaussian score.
"basis_function_blocks_bic" - BIC score for mixed data using basis-function models.
"basis_function_sem_bic" - SEM BIC score for basis-function models.
"conditional_gaussian" - Mixed discrete/continuous BIC score.
"degenerate_gaussian" - Degenerate Gaussian BIC score.
"discrete_bic" - BIC score for discrete data.
"ebic" - Extended BIC score.
"gic" - Generalized Information Criterion (GIC) score.
"mag_degenerate_gaussian_bic" - MAG Degenerate Gaussian BIC Score.
"poisson_prior" - Poisson prior score.
"rank_bic" - Rank-based BIC score.
"sem_bic" - SEM BIC score.
"zhang_shen_bound" - Zhang and Shen bound score.

...

Additional arguments passed to the private score-setting methods. For the following scores, the following parameters are available:

"bdeu" - Bayes Dirichlet Equivalent score with uniform priors.
- sample_prior = 10 - This sets the prior equivalent sample size. This number is added to the sample size for each conditional probability table in the model and is divided equally among the cells in the table,
- singularity_lambda = 0.0 - Small number >= 0: Add lambda to the diagonal, < 0 Pseudoinverse.
"basis_function_bic" - BIC score for basis-function models. This is a generalization of the Degenerate Gaussian score.
- truncation_limit = 3 - Basis functions 1 though this number will be used. The Degenerate Gaussian category indicator variables for mixed data are also used,
- penalty_discount = 2 - Penalty discount. Higher penalty yields sparser graphs,
- singularity_lambda = 0.0 - Small number >= 0: Add lambda to the diagonal, < 0 Pseudoinverse,
- do_one_equation_only = FALSE - If TRUE, only one equation should be used when expanding the basis.
"basis_function_blocks_bic" - BIC score for mixed data using basis-function models.
- basis_type = "polynomial" - The type of basis to use. Supported types are "polynomial", "legendre", "hermite", and "chebyshev",
- penalty_discount = 2 - Penalty discount factor used in BIC = 2L - ck log N, where c is the penalty. Higher c yield sparser graphs,
- truncation_limit = 3 - Basis functions 1 through this number will be used. The Degenerate Gaussian category indicator variables for mixed data are also used.
"basis_function_sem_bic" - SEM BIC score for basis-function models.
- penalty_discount = 2 - Penalty discount factor used in BIC = 2L - ck log N, where c is the penalty. Higher c yield sparser graphs,
- jitter = 1e-8 - Small non-negative constant added to the diagonal of covariance/correlation matrices for numerical stability,
- truncation_limit = 3 - Basis functions 1 through this number will be used. The Degenerate Gaussian category indicator variables for mixed data are also used.
"conditional_gaussian" - Mixed discrete/continuous BIC score.
- penalty_discount = 1 - Penalty discount. Higher penalty yields sparser graphs,
- discretize = TRUE - If TRUE for the conditional Gaussian likelihood, when scoring X –> D where X is continuous and D discrete, one should to simply discretize X for just those cases. If FALSE, the integration will be exact,
- num_categories_to_discretize = 3 - In case the exact algorithm is not used for discrete children and continuous parents is not used, this parameter gives the number of categories to use for this second (discretized) backup copy of the continuous variables,
- structure_prior = 0 - The default number of parents for any conditional probability table. Higher weight is accorded to tables with about that number of parents. The prior structure weights are distributed according to a binomial distribution.
"degenerate_gaussian" - Degenerate Gaussian BIC score.
- penalty_discount = 1 - Penalty discount. Higher penalty yields sparser graphs,
- structure_prior = 0 - The default number of parents for any conditional probability table. Higher weight is accorded to tables with about that number of parents. The prior structure weights are distributed according to a binomial distribution,
- singularity_lambda = 0.0 - Small number >= 0: Add lambda to the diagonal, < 0 Pseudoinverse.
"discrete_bic" - BIC score for discrete data.
- penalty_discount = 2 - Penalty discount. Higher penalty yields sparser graphs,
- structure_prior = 0 - The default number of parents for any conditional probability table. Higher weight is accorded to tables with about that number of parents. The prior structure weights are distributed according to a binomial distribution.
"ebic" - Extended BIC score.
- gamma - The gamma parameter in the EBIC score.
- singularity_lambda = 0.0 - Small number >= 0: Add lambda to the diagonal, < 0 Pseudoinverse.
"gic" - Generalized Information Criterion (GIC) score.
- penalty_discount = 1 - Penalty discount. Higher penalty yields sparser graphs,
- sem_gic_rule = "bic" - The following rules are available: "bic" - $\ln n$, "gic2" - $p n^{1/3}$, "ric" - $2 \ln(p n)$, "ricc" - $2(\ln(p n) + \ln\ln(p n))$, "gic6" - $\ln n \ln(p n)$.
- singularity_lambda = 0.0 - Small number >= 0: Add lambda to the diagonal, < 0 Pseudoinverse.
"mag_degenerate_gaussian_bic" - MAG Degenerate Gaussian BIC Score.
- penalty_discount = 1 - Penalty discount. Higher penalty yields sparser graphs,
- structure_prior = 0 - The default number of parents for any conditional probability table. Higher weight is accorded to tables with about that number of parents. The prior structure weights are distributed according to a binomial distribution,
"poisson_prior" - Poisson prior score.
- poisson_lambda = 1 - Lambda parameter for the Poisson distribution (> 0),
- singularity_lambda = 0.0 - Small number >= 0: Add lambda to the diagonal, < 0 Pseudoinverse.
"sem_bic" - SEM BIC score.
- penalty_discount = 2 - Penalty discount factor used in BIC = 2L - ck log N, where c is the penalty. Higher c yield sparser graphs,
- structure_prior = 0 - The default number of parents for any conditional probability table. Higher weight is accorded to tables with about that number of parents. The prior structure weights are distributed according to a binomial distribution,
- singularity_lambda = 0.0 - Small number >= 0: Add lambda to the diagonal, < 0 Pseudoinverse.
"rank_bic" - Rank-based BIC score.
- gamma = 0.8 - Gamma parameter for Extended BIC (Chen and Chen, 2008). Between 0 and 1,
- penalty_discount = 2 - Penalty discount factor used in BIC = 2L - ck log N, where c is the penalty. Higher c yield sparser graphs.
"zhang_shen_bound" - Zhang and Shen bound score.
- risk_bound = 0.2 - This is the probability of getting the true model if a correct model is discovered. Could underfit.
- singularity_lambda = 0.0 - Small number >= 0: Add lambda to the diagonal, < 0 Pseudoinverse.

Returns

Invisibly returns self.

Method `set_alg()`

Sets the causal discovery algorithm to use in Tetrad.

Usage

TetradSearch$set_alg(method, ...)

Arguments

method

(character) Name of the algorithm (e.g., "fges", "pc", "fci", etc.).

...

Additional parameters passed to the private algorithm-setting methods. For the following algorithms, the following parameters are available:

"boss" - BOSS algorithm.
- num_starts = 1 - The number of times the algorithm should be started from different initializations. By default, the algorithm will be run through at least once using the initialized parameters,
- use_bes = TRUE - If TRUE, the algorithm uses the backward equivalence search from the GES algorithm as one of its steps,
- use_data_order = TRUE - If TRUE, the data variable order should be used for the first initial permutation,
- output_cpdag = TRUE - If TRUE, the DAG output of the algorithm is converted to a CPDAG.
"boss_fci" - BOSS-FCI algorithm.
- depth = -1 - Maximum size of conditioning set, Set to -1 for unlimited,
- max_disc_path_length = -1 - Maximum length for any discriminating path, Set to -1 for unlimited,
- use_bes = TRUE - If TRUE, the algorithm uses the backward equivalence search from the GES algorithm as one of its steps,
- use_heuristic = FALSE - If TRUE, use the max p heuristic version,
- complete_rule_set_used = TRUE - FALSE if the (simpler) final orientation rules set due to P. Spirtes, guaranteeing arrow completeness, should be used; TRUE if the (fuller) set due to J. Zhang, should be used guaranteeing additional tail completeness,
- guarantee_pag = FALSE - Ensure the output is a legal PAG (where feasible).
"fci" - FCI algorithm.
- depth = -1 - Maximum size of conditioning set,
- stable_fas = TRUE - If TRUE, the "stable" version of the PC adjacency search is used, which for k > 0 fixes the graph for depth k + 1 to that of the previous depth k.
- max_disc_path_length = -1 - Maximum length for any discriminating path,
- complete_rule_set_used = TRUE - FALSE if the (simpler) final orientation rules set due to P. Spirtes, guaranteeing arrow completeness, should be used; TRUE if the (fuller) set due to J. Zhang, should be used guaranteeing additional tail completeness.
- guarantee_pag = FALSE - Ensure the output is a legal PAG (where feasible).
"ges" ("fges") - Fast Greedy Equivalence Search (FGES) algorithm.
- symmetric_first_step = FALSE - If TRUE, scores for both X –> Y and X <– Y will be calculated and the higher score used.
- max_degree = -1 - Maximum degree of any node in the graph. Set to -1 for unlimited,
- parallelized = FALSE - If TRUE, the algorithm should be parallelized,
- faithfulness_assumed = FALSE - If TRUE, assume that if $X \perp\!\!\!\perp Y$ (by an independence test) then $X \perp\!\!\!\perp Y$ | Z for nonempty Z.
"gfci" - GFCI algorithm. Combines FGES and FCI.
- depth = -1 - Maximum size of conditioning set,
- max_degree = -1 - Maximum degree of any node in the graph. Set to -1 for unlimited,
- max_disc_path_length = -1 - Maximum length for any discriminating path,
- complete_rule_set_used = TRUE - FALSE if the (simpler) final orientation rules set due to P. Spirtes, guaranteeing arrow completeness, should be used; TRUE if the (fuller) set due to J. Zhang, should be used guaranteeing additional tail completeness,
- guarantee_pag = FALSE - Ensure the output is a legal PAG (where feasible),
- use_heuristic = FALSE - If TRUE, use the max p heuristic.
- start_complete = FALSE - If TRUE, start from a complete graph.
"grasp" - GRaSP (Greedy Relations of Sparsest Permutation) algorithm.
- covered_depth = 4 - The depth of recursion for first search,
- singular_depth = 1 - Recursion depth for singular tucks,
- nonsingular_depth = 1 - Recursion depth for nonsingular tucks,
- ordered_alg = FALSE - If TRUE, earlier GRaSP stages should be performed before later stages,
- raskutti_uhler = FALSE - If TRUE, use Raskutti and Uhler's DAG-building method (test); if FALSE, use Grow-Shrink (score).
- use_data_order = TRUE - If TRUE, the data variable order should be used for the first initial permutation,
- num_starts = 1 - The number of times the algorithm should be started from different initializations. By default, the algorithm will be run through at least once using the initialized parameters.
"grasp_fci" - GRaSP-FCI algorithm. Combines GRaSP and FCI.
- depth = -1 - Maximum size of conditioning set,
- stable_fas = TRUE - If TRUE, the "stable" version of the PC adjacency search is used, which for k > 0 fixes the graph for depth k + 1 to that of the previous depth k.
- max_disc_path_length = -1 - Maximum length for any discriminating path,
- complete_rule_set_used = TRUE - FALSE if the (simpler) final orientation rules set due to P. Spirtes, guaranteeing arrow completeness, should be used; TRUE if the (fuller) set due to J. Zhang, should be used guaranteeing additional tail completeness,
- covered_depth = 4 - The depth of recursion for first search,
- singular_depth = 1 - Recursion depth for singular tucks,
- nonsingular_depth = 1 - Recursion depth for nonsingular tucks,
- ordered_alg = FALSE - If TRUE, earlier GRaSP stages should be performed before later stages,
- raskutti_uhler = FALSE - If TRUE, use Raskutti and Uhler's DAG-building method (test); if FALSE, use Grow-Shrink (score).
- use_data_order = TRUE - If TRUE, the data variable order should be used for the first initial permutation,
- num_starts = 1 - The number of times the algorithm should be started from different initializations. By default, the algorithm will be run through at least once using the initialized parameters,
- guarantee_pag = FALSE - If TRUE, ensure the output is a legal PAG (where feasible).
"pc" - Peter-Clark (PC) algorithm
- conflict_rule = 1 - The value of conflict_rule determines how collider conflicts are handled. 1 corresponds to the "overwrite" rule as introduced in the pcalg package, see pcalg::pc(). 2 means that all collider conflicts using bidirected edges should be prioritized, while 3 means that existing colliders should be prioritized, ignoring subsequent conflicting information.
- depth = -1 - Maximum size of conditioning set,
- stable_fas = TRUE - If TRUE, the "stable" version of the PC adjacency search is used, which for k > 0 fixes the graph for depth k + 1 to that of the previous depth k.
- guarantee_cpdag = FALSE - If TRUE, ensure the output is a legal CPDAG.
"rfci" - Restricted FCI algorithm
- depth = -1 - Maximum size of conditioning set,
- stable_fas = TRUE - If TRUE, the "stable" version of the PC adjacency search is used, which for k > 0 fixes the graph for depth k + 1 to that of the previous depth k.
- max_disc_path_length = -1 - Maximum length for any discriminating path,
- complete_rule_set_used = TRUE - FALSE if the (simpler) final orientation rules set due to P. Spirtes, guaranteeing arrow completeness, should be used; TRUE if the (fuller) set due to J. Zhang, should be used guaranteeing additional tail completeness.
- guarantee_pag = FALSE - Ensure the output is a legal PAG (where feasible).
"sp_fci" - Sparsest Permutation using FCI
- depth = -1 - Maximum size of conditioning set,
- max_disc_path_length = -1 - Maximum length for any discriminating path,
- complete_rule_set_used = TRUE - FALSE if the (simpler) final orientation rules set due to P. Spirtes, guaranteeing arrow completeness, should be used; TRUE if the (fuller) set due to J. Zhang, should be used guaranteeing additional tail completeness,
- guarantee_pag = FALSE - Ensure the output is a legal PAG (where feasible),
- use_heuristic = FALSE - If TRUE, use the max p heuristic version.

Returns

Invisibly returns self.

Method `set_knowledge()`

Sets the background Knowledge object.

Usage

TetradSearch$set_knowledge(knowledge_obj)

Arguments

knowledge_obj: An object containing Tetrad knowledge.

Method `set_params()`

Sets parameters for the Tetrad search.

Usage

TetradSearch$set_params(...)

Arguments

...: Named arguments for the parameters to set.

Method `get_parameters_for_function()`

Retrieves the argument names of a matching private function.

Usage

TetradSearch$get_parameters_for_function(
  fn_pattern,
  score = FALSE,
  test = FALSE,
  alg = FALSE
)

Arguments

fn_pattern: (character) A pattern that should match a private method name.
score: If TRUE, retrieves parameters for a scoring function.
test: If TRUE, retrieves parameters for a test function.
alg: If TRUE, retrieves parameters for an algorithm.

Returns

(character) The names of the parameters.

Method `run_search()`

Runs the chosen Tetrad algorithm on the data.

Usage

TetradSearch$run_search(
  data = NULL,
  bootstrap = FALSE,
  int_cols_as_cont = TRUE
)

Arguments

data: (optional) If provided, overrides the previously set data.
bootstrap: (logical) If TRUE, bootstrapped graphs will be generated.
int_cols_as_cont: (logical) If TRUE, integer columns are treated as continuous, since Tetrad does not support ordinal data, but only either continuous or nominal data. Default is TRUE.

Returns

A Disco object (a list with a caugi and a Knowledge object). Also populates self$java.

Method `set_bootstrapping()`

Configures bootstrapping parameters for the Tetrad search.

Usage

TetradSearch$set_bootstrapping(
  number_resampling = 0,
  percent_resample_size = 100,
  add_original = TRUE,
  with_replacement = TRUE,
  resampling_ensemble = 1,
  seed = -1
)

Arguments

number_resampling: (integer) Number of bootstrap samples.
percent_resample_size: (numeric) Percentage of sample size for each bootstrap.
add_original: (logical) If TRUE, add the original dataset to the bootstrap set.
with_replacement: (logical) If TRUE, sampling is done with replacement.
resampling_ensemble: (integer) How the resamples are used or aggregated.
seed: (integer) Random seed, or -1 for none.

Method `set_data()`

Sets or overrides the data used by Tetrad.

Usage

TetradSearch$set_data(data, int_cols_as_cont = TRUE)

Arguments

data: (data.frame) The new data to load.
int_cols_as_cont: (logical) If TRUE, integer columns are treated as continuous, since Tetrad does not support ordinal data, but only either continuous or nominal data. Default is TRUE.

Method `set_verbose()`

Toggles the verbosity in Tetrad.

Usage

TetradSearch$set_verbose(verbose)

Arguments

verbose: (logical) TRUE to enable verbose logging, FALSE otherwise.

Method `set_time_lag()`

Sets an integer time lag for time-series algorithms.

Usage

TetradSearch$set_time_lag(time_lag = 0)

Arguments

time_lag: (integer) The time lag to set.

Method `get_data()`

Retrieves the current Java data object.

Usage

TetradSearch$get_data()

Returns

(Java object) Tetrad dataset.

Method `get_knowledge()`

Returns the background Knowledge object.

Usage

TetradSearch$get_knowledge()

Returns

(Java object) Tetrad Knowledge.

Method `get_java()`

Gets the main Java result object (usually a graph) from the last search.

Usage

TetradSearch$get_java()

Returns

(Java object) The Tetrad result graph or model.

Method `get_string()`

Returns the string representation of a given Java object or self$java.

Usage

TetradSearch$get_string(java_obj = NULL)

Arguments

java_obj: (Java object, optional) If NULL, uses self$java.

Returns

(character) The toString() of that Java object.

Method `get_dot()`

Produces a DOT (Graphviz) representation of the graph.

Usage

TetradSearch$get_dot(java_obj = NULL)

Arguments

java_obj: (Java object, optional) If NULL, uses self$java.

Returns

(character) The DOT-format string.

Method `get_amat()`

Produces an amat representation of the graph.

Usage

TetradSearch$get_amat(java_obj = NULL)

Arguments

java_obj: (Java object, optional) If NULL, uses self$java.

Returns

(character) The adjacency matrix.

Method `clone()`

The objects of this class are cloneable with this method.

Usage

TetradSearch$clone(deep = FALSE)

Arguments

deep: Whether to make a deep clone.

Examples

### tetrad_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.

# Requires Tetrad to be installed
if (verify_tetrad()$installed && verify_tetrad()$java_ok) {
  data(num_data)

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

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

  # Example with detailed settings:
  my_pc2 <- pc(
    engine = "tetrad",
    test = "sem_bic",
    penalty_discount = 1,
    structure_prior = 1,
    singularity_lambda = 0.1
  )
  disco(data = num_data, method = my_pc2)

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

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

  g <- s$run_search()

  print(g)
}
#> 
#> ── caugi graph ─────────────────────────────────────────────────────────────────
#> Graph class: UNKNOWN
#> 
#> ── Edges ──
#> 
#>   from  edge  to   
#>   <chr> <chr> <chr>
#> 1 X1    -->   Y    
#> 2 X1    ---   Z    
#> 3 X2    ---   X3   
#> 4 X2    -->   Y    
#> 5 X3    -->   Y    
#> 6 Z     -->   Y    
#> ── Nodes ──
#> 
#>   name 
#>   <chr>
#> 1 X1   
#> 2 X2   
#> 3 X3   
#> 4 Z    
#> 5 Y    
#> ── Knowledge object ────────────────────────────────────────────────────────────

Public fields

Methods

Public methods

Method new()

Usage

Method set_test()

Usage

Arguments

Returns

Method set_score()

Usage

Arguments

Returns

Method set_alg()

Usage

Arguments

Returns

Method set_knowledge()

Usage

Arguments

Method set_params()

Usage

Arguments

Method get_parameters_for_function()

Usage

Arguments

Returns

Method run_search()

Usage

Arguments

Returns

Method set_bootstrapping()

Usage

Arguments

Method set_data()

Usage

Arguments

Method set_verbose()

Usage

Arguments

Method set_time_lag()

Usage

Arguments

Method get_data()

Usage

Returns

Method get_knowledge()

Usage

Returns

Method get_java()

Usage

Returns

Method get_string()

Usage

Arguments

Returns

Method get_dot()

Usage

Arguments

Returns

Method get_amat()

Usage

Arguments

Returns

Method clone()

Usage

Arguments

Examples

Method `new()`

Method `set_test()`

Method `set_score()`

Method `set_alg()`

Method `set_knowledge()`

Method `set_params()`

Method `get_parameters_for_function()`

Method `run_search()`

Method `set_bootstrapping()`

Method `set_data()`

Method `set_verbose()`

Method `set_time_lag()`

Method `get_data()`

Method `get_knowledge()`

Method `get_java()`

Method `get_string()`

Method `get_dot()`

Method `get_amat()`

Method `clone()`