Class Kpc
java.lang.Object
edu.cmu.tetrad.search.work_in_progress.Kpc
- All Implemented Interfaces:
IGraphSearch
Kernelized PC algorithm. This is the same as the PC class, the nonparametric kernel-based HSIC test is used for independence testing and the parameters for this test can be set directly when Kpc is initialized.
Moving this to the work_in_progress package because it has not been tested in a very long time, and there is another option available that has been tested, namely, to run PC using the KCI test due to Kun Zhang.
- Author:
- Robert Tillman.
-
Constructor Summary
Constructors -
Method Summary
Modifier and TypeMethodDescriptiondouble
getAlpha()
Gets the getModel significance level.int
getDepth()
long
int
getPerms()
Gets the getModel number of bootstrap samples useddouble
Gets the getModel precision for the Incomplete Choleskyboolean
search()
Runs PC starting with a complete graph over all nodes of the given conditional independence test, using the given independence test and knowledge and returns the resultant graph.Runs PC starting with a commplete graph over the given list of nodes, using the given independence test and knowledge and returns the resultant graph.void
setAlpha
(double alpha) Sets the significance level at which independence judgments should be made.void
setDepth
(int depth) Sets the depth of the search--that is, the maximum number of conditioning nodes for any conditional independence checked.void
setKnowledge
(Knowledge knowledge) Sets the knowledge specification to be used in the search.void
setMeekPreventCycles
(boolean meekPreventCycles) void
setPerms
(int perms) Set the number of bootstrap samples to usevoid
setVerbose
(boolean verbose) Sets whether verbose output should be printed.
-
Constructor Details
-
Kpc
Constructs a new PC search using for the given dataset.- Parameters:
dataset
- The oracle for conditional independence facts. This does not make a copy of the independence test, for fear of duplicating the data set!
-
-
Method Details
-
isMeekPreventCycles
public boolean isMeekPreventCycles()- Returns:
- true iff edges will not be added if they would create cycles.
-
setMeekPreventCycles
public void setMeekPreventCycles(boolean meekPreventCycles) - Parameters:
meekPreventCycles
- Set to true just in case edges will not be addeds if they would create cycles.
-
getIndependenceTest
- Returns:
- the independence test being used in the search.
-
getKnowledge
- Returns:
- the knowledge specification used in the search. Non-null.
-
setKnowledge
Sets the knowledge specification to be used in the search. May not be null. -
getSepset
- Returns:
- the sepset map from the most recent search. Non-null after the first call to
search()
.
-
getDepth
public int getDepth()- Returns:
- the getModel depth of search--that is, the maximum number of conditioning nodes for any conditional independence checked.
-
setDepth
public void setDepth(int depth) Sets the depth of the search--that is, the maximum number of conditioning nodes for any conditional independence checked.- Parameters:
depth
- The depth of the search. The default is 1000. A value of -1 may be used to indicate that the depth should be high (1000). A value of Integer.MAX_VALUE may not be used, due to a bug on multi-core machines.
-
search
Runs PC starting with a complete graph over all nodes of the given conditional independence test, using the given independence test and knowledge and returns the resultant graph. The returned graph will be a CPDAG if the independence information is consistent with the hypothesis that there are no latent common causes. It may, however, contain cycles or bidirected edges if this assumption is not born out, either due to the actual presence of latent common causes, or due to statistical errors in conditional independence judgments.- Specified by:
search
in interfaceIGraphSearch
- Returns:
- The discovered graph.
-
search
Runs PC starting with a commplete graph over the given list of nodes, using the given independence test and knowledge and returns the resultant graph. The returned graph will be a CPDAG if the independence information is consistent with the hypothesis that there are no latent common causes. It may, however, contain cycles or bidirected edges if this assumption is not born out, either due to the actual presence of latent common causes, or due to statistical errors in conditional independence judgments.All of the given nodes must be in the domain of the given conditional independence test.
-
getElapsedTime
public long getElapsedTime()- Returns:
- the elapsed time of the search, in milliseconds.
-
getUnshieldedColliders
- Returns:
- the set of unshielded colliders in the graph returned by
search()
. Non-null aftersearch
is called.
-
getUnshieldedNoncolliders
- Returns:
- the set of unshielded noncolliders in the graph returned by
search()
. Non-null aftersearch
is called.
-
setVerbose
public void setVerbose(boolean verbose) Sets whether verbose output should be printed.- Parameters:
verbose
- True, if so.
-
getAlpha
public double getAlpha()Gets the getModel significance level. -
setAlpha
public void setAlpha(double alpha) Sets the significance level at which independence judgments should be made. -
getPrecision
public double getPrecision()Gets the getModel precision for the Incomplete Cholesky -
getPerms
public int getPerms()Gets the getModel number of bootstrap samples used -
setPerms
public void setPerms(int perms) Set the number of bootstrap samples to use
-