parametric
fun parametricComponentCountFrequency(fitted: ContinuousDistributionIfc, sampleSize: Int, numBootstrapSamples: Int = defaultNumBootstrapSamples, numComponentsRange: IntRange = MixtureModeler.defaultNumComponentsRange, fitter: () -> ComponentFitterIfc = { PDFComponentFitter() }, criterion: MixtureCriterionIfc = MixtureBICCriterion(), refiner: () -> PartitionRefinerIfc = MixtureModeler.defaultRefiner, selector: () -> FamilySelectorIfc = MixtureModeler.defaultSelector, streamNum: Int = 0, streamProvider: RNStreamProviderIfc = KSLRandom.DefaultRNStreamProvider): MixtureBootstrapResults(source)
Refits the mixture on samples drawn from a fitted mixture rather than from the data.
Produces no repeated values, so the tie interaction described above does not arise. In exchange it assumes the supplied mixture is the truth, so it measures how well the procedure recovers a known answer rather than how much the data constrain it. Read it as the optimistic bound.
Parameters
fitted
the mixture to draw from, usually a previous recommendation
sample
the size of each drawn sample, usually the original sample size