dMixmod: Internal S4 classes.

In Infusion: Inference Using Simulation

Internal S4 classes.

Description

The objects or methods referenced here are not to be called by the user, or are waiting for documentation to be written.

dMixmod is an S4 class describing some distributions that extend the multivariate gaussian mixture models (MGMM) by possibly involving discrete probability masses for some variables and gaussian mixtures for other variables conditional on such discrete events. In terms of the represented probability models, and of its slots, is effectively extends the MixmodResults class from the Rmixmod package. But it does not formally extends this class in terms of OOP programming. It should not be considered as part of the programming interface, and may be subject to backward-incompatible modifications without notice. In the current implementation it cannot represent general mixtures of discrete probabilities and MGMMs, and may yield correct results only for the degenerate case of pure MGMMs or when inference can be based on the conditional density of continuous variables conditional on the (joint-, if relevant) discrete event observed in the data.

Usage

# dMixmod: Don't try to use it! It's for programming only.

Value

A dMixmod object has the same slots as a MixmodResults object, plus additional ones: @freq is the frequency of the conditioning event for the gaussian mixture model. In the Infusion code, this event is defined jointly by the “observed” summary statistics and the reference simulation table: a probability mass for specific values v is identified from the simulated distribution of summary statistics in the reference table, and freq is an estimate of the probability mass if the summary statistics match v, or the converse probability if they do not match.

Note

Use str(attributes(.)) to see the slots of a dMixmod object if str(.) does not work.

Examples

# The dMixmod object can be used internally to handle repeated and boundary values 
# of summary statistics. The user has to add an attribute to the observations,
# as explained in help("boundaries-attribute"):
Sobs <- c(mean=4.321, se=0.987) # hypothetical observation
attr(Sobs,"boundaries") <- c(someSummStat=-1)  

Infusion documentation built on May 3, 2023, 5:10 p.m.