Package edu.cmu.tetrad.search.unmix

Class UnmixDiagnostics.EntropyStats

java.lang.Object

edu.cmu.tetrad.search.unmix.UnmixDiagnostics.EntropyStats

Enclosing class:

UnmixDiagnostics

public static final class UnmixDiagnostics.EntropyStats
extends Object

Represents the entropy statistics of clustering results in the context of mixture models. This class provides metrics to evaluate the uncertainty and confidence in clustering assignments based on the responsibility matrix.

Field Summary

Fields

Modifier and Type	Field	Description
final double	fracConf80	Represents the fraction of data points in the clustering result where the maximum responsibility value assigned to a cluster for a data point (r_ik) is greater than or equal to 0.80.
final double	fracConf90	Represents the fraction of data points in the clustering result where the maximum responsibility value assigned to a cluster for a data point (r_ik) is greater than or equal to 0.90.
final double	meanEntropy	Represents the mean entropy of clustering results, providing a measure of the uncertainty in cluster assignments.

Constructor Summary

Constructors

Constructor	Description
EntropyStats(double meanEntropy, double fracConf90, double fracConf80)	Constructs an instance of EntropyStats with the specified mean entropy and confidence fractions.

Method Summary

Methods inherited from class java.lang.Object

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

Field Details

meanEntropy

public final double meanEntropy

Represents the mean entropy of clustering results, providing a measure of the uncertainty in cluster assignments. The value ranges between 0 and 1: - A value of 0 indicates a "crisp" clustering where each data point is assigned to a single cluster with high certainty. - A value of 1 indicates a high level of uncertainty, with a "uniform" distribution of responsibility values across clusters.

fracConf90

public final double fracConf90

Represents the fraction of data points in the clustering result where the maximum responsibility value assigned to a cluster for a data point (r_ik) is greater than or equal to 0.90. This metric provides an indication of the proportion of data points that are assigned to a cluster with high confidence.

fracConf80

public final double fracConf80

Represents the fraction of data points in the clustering result where the maximum responsibility value assigned to a cluster for a data point (r_ik) is greater than or equal to 0.80. This metric provides an indication of the proportion of data points that are assigned to a cluster with moderate confidence.

Constructor Details

EntropyStats

public EntropyStats(double meanEntropy,
 double fracConf90,
 double fracConf80)

Constructs an instance of EntropyStats with the specified mean entropy and confidence fractions.

Parameters:

meanEntropy - the mean entropy of clustering results, representing the average uncertainty in cluster assignments. Values range from 0 (crisp clustering) to 1 (maximum uncertainty).

fracConf90 - the fraction of data points where the maximum responsibility value for a cluster is greater than or equal to 0.90, indicating the proportion of highly confident cluster assignments.

fracConf80 - the fraction of data points where the maximum responsibility value for a cluster is greater than or equal to 0.80, indicating the proportion of moderately confident cluster assignments.