KSLCore/ksl.utilities.io.report.extensions/multipleComparison

multipleComparison

fun ReportBuilder.multipleComparison(mca: MultipleComparisonAnalyzer, direction: MCBDirection = MCBDirection.BOTH, indifferenceZone: Double = mca.defaultIndifferenceZone, altConfidenceLevel: Double = mca.defaultLevel, diffConfidenceLevel: Double = mca.defaultLevel, probCorrectSelection: Double = mca.defaultLevel, showAltCIPlot: Boolean = false, showBoxPlot: Boolean = false, xAxisLabel: String? = null, yAxisLabel: String? = null)(source)

Appends a self-contained section for a MultipleComparisonAnalyzer.

Produces (inside a section titled mca.name or "Multiple Comparison Analysis" when the name is blank):

Alternative Statistics — StatTable at altConfidenceLevel
Pairwise Differences — DataTable of half-width CIs at diffConfidenceLevel
MCB Max Intervals (when direction is MCBDirection.MAX or MCBDirection.BOTH) — DataTable with columns Alternative | Lower | Upper | Possible Best and a ConfidenceIntervalsPlot. An alternative is Possible Best for max when its upper interval limit is strictly greater than zero.
MCB Min Intervals (when direction is MCBDirection.MIN or MCBDirection.BOTH) — same structure. An alternative is Possible Best for min when its lower interval limit is strictly less than zero.
Screening — direction-matched DataTable(s) of alternatives surviving screening at probCorrectSelection.

Usage:

// Full report — both max and min, all defaults from the MCA object
val doc = report("Throughput Study") {
    multipleComparison(mca)
}

// Minimum only, custom indifference zone and PCS
val doc = report("Cost Study") {
    multipleComparison(
        mca,
        direction           = MCBDirection.MIN,
        indifferenceZone    = 5.0,
        probCorrectSelection = 0.90
    )
}

Parameters

mca

the multiple comparison analyzer to report

direction

which MCB direction(s) to render; default MCBDirection.BOTH

indifferenceZone

delta for MCB interval construction; defaults to MultipleComparisonAnalyzer.defaultIndifferenceZone

altConfidenceLevel

confidence level for the alternatives StatTable; defaults to mca.defaultLevel

diffConfidenceLevel

confidence level for the pairwise differences table; defaults to mca.defaultLevel

probCorrectSelection

probability of correct selection for screening; defaults to mca.defaultLevel

showAltCIPlot

true inserts an "Alternative Confidence Intervals" sub-section immediately after the Alternative Statistics table; the section contains a ConfidenceIntervalsPlot showing the across-replication CI for each alternative's mean at altConfidenceLevel; defaults to false

showBoxPlot

true inserts a "Response Distributions" sub-section after the CI plot sub-section (or directly after the statistics table when showAltCIPlot is false); the section contains a ksl.utilities.io.plotting.MultiBoxPlot with one box per alternative sourced from MultipleComparisonAnalyzer.observationsAsMap; defaults to false

xAxisLabel

optional override for the x-axis label of every plot emitted by this section (alt-CI, box plot, MCB max, MCB min). Blank / null falls back to "Alternative".

yAxisLabel

optional override for the y-axis label of every plot emitted by this section. Blank / null falls back to the MCA's response name (MultipleComparisonAnalyzer.name) or "Value" when the name is blank.