KSLCore/ksl.utilities.distributions.fitting.mixture.scoring

Package-level declarations

Types

class HeldOutLogLikelihoodCriterion(holdOutData: DoubleArray, worstValue: Double = LogLikelihoodCriterion.defaultWorstValue) : MixtureCriterionIfc

The mean log-likelihood of a fitted mixture on observations it was not fitted to.

abstract class LogLikelihoodCriterion(val name: String, val worstValue: Double = defaultWorstValue) : MixtureCriterionIfc

Shared machinery for criteria built from the observed-data log-likelihood.

The Akaike information criterion of a fitted mixture, 2 * p - 2 * logLikelihood.

The Bayesian information criterion of a fitted mixture.

A criterion by which fitted mixtures are ranked.

data class MixtureCriterionValue(val value: Double, val isComparable: Boolean, val numZeroDensity: Int, val message: String? = null)

The value of a criterion for one candidate, together with whether that value may be compared with another candidate's.

class MixtureEBICCriterion(val catalogSize: Int, val xi: Double = defaultXi, worstValue: Double = defaultWorstValue) : LogLikelihoodCriterion

The Bayesian information criterion with an additional penalty for having selected the component families from a catalog.

The Hannan-Quinn information criterion of a fitted mixture, -2 * logLikelihood + 2 * p * ln(ln(n)).

The integrated completed likelihood, in its approximation by the Bayesian information criterion penalized by the classification entropy.