Mixture
data class MixtureLogLikelihoodResult(val value: Double, val numZeroDensity: Int, val numResponsibilityUndefined: Int)(source)
The observed-data log-likelihood of a mixture, together with the diagnostics needed to know whether that number can be compared with another one.
Parameters
value
the sum over observations of the log of the mixture density. Negative infinity when at least one observation lies where every component has zero density.
num
the number of observations at which the mixture density is zero. A candidate with any such observation is not comparable on an information criterion with one that has none, so this count is reported rather than absorbed.
num
the number of observations at which the mixture density is zero and therefore no component responsibility is defined. This is a model-state failure in the sense of Proposition 5(a), not a zero responsibility for every component.