MeanEstimate

data class MeanEstimate(val mean: Double, val effectiveSamples: Double, val responseVariance: Double, val knownNoiseVariance: Double?, val varianceOfMean: Double?)(source)

Weighted mean with separate response dispersion and uncertainty in its estimate.

Constructors

constructor(mean: Double, effectiveSamples: Double, responseVariance: Double, knownNoiseVariance: Double?, varianceOfMean: Double?)(source)

Kish effective sample size; invariant under a common rescaling of weights.

Weighted average supplied observation variance, null when none is supplied.

Weighted response mean.

Unbiased weighted response dispersion, or zero when it cannot be estimated.

Standard error, infinite when a single unknown-noise observation cannot estimate it.

Maximum of empirical sampling variance and propagated known noise; null without enough evidence.

Descriptive mean minus a standard-error multiple; not a sequential confidence bound.

Descriptive mean plus a standard-error multiple; not a sequential confidence bound.