Package edu.cmu.tetrad.search.blocks

Class FtfcBlockDiscoverer

java.lang.Object

edu.cmu.tetrad.search.blocks.FtfcBlockDiscoverer

All Implemented Interfaces:

BlockDiscoverer

public class FtfcBlockDiscoverer
extends Object
implements BlockDiscoverer

A concrete implementation of the BlockDiscoverer interface that discovers clusters or "blocks" of indices using the FTFC algorithm. The discovered blocks are represented as a BlockSpec and include features such as validation, canonicalization, and policy-based adjustments.

This class leverages the FOFC algorithm to generate clusters by analyzing the provided dataset and utilizing a statistical test with a specified confidence level and equivalent sample size. The discovered clusters can be further refined based on policy configurations for handling overlapping or conflicting blocks.

Constructor Summary

Constructors

Constructor	Description
FtfcBlockDiscoverer(DataSet dataSet, double alpha, int ess, SingleClusterPolicy policy, boolean verbose)	Constructs a new instance of FofcBlockDiscoverer, which is used to discover clusters or "blocks" of indices in a dataset based on the FOFC algorithm.

Method Summary

Modifier and Type	Method	Description
BlockSpec	discover()	Discovers clusters or "blocks" of indices in the dataset using the FOFC algorithm.

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

FtfcBlockDiscoverer

public FtfcBlockDiscoverer(DataSet dataSet,
 double alpha,
 int ess,
 SingleClusterPolicy policy,
 boolean verbose)

Constructs a new instance of FofcBlockDiscoverer, which is used to discover clusters or "blocks" of indices in a dataset based on the FOFC algorithm. The discovered blocks are adjusted and refined according to the specified configuration parameters.

Parameters:

dataSet - the dataset to be analyzed for block discovery.

alpha - the significance level used in the statistical test to determine independence.

ess - the equivalent sample size parameter used in Bayesian methods within the FOFC algorithm.

policy - the policy to handle scenarios involving overlapping or conflicting blocks.

verbose - whether to enable verbose output during the block discovery process.

Method Details

discover

public BlockSpec discover()

Discovers clusters or "blocks" of indices in the dataset using the FOFC algorithm. The method performs block validation, canonicalization, and policy-based adjustments to refine the discovered clusters.

Specified by:

discover in interface BlockDiscoverer

Returns:

a BlockSpec object representing the discovered and refined set of blocks.