toComparisonReport

fun KSLDatabase.toComparisonReport(responseName: String, expNames: List<String>, title: String = "Comparison — ", direction: MCBDirection = MCBDirection.BOTH, indifferenceZone: Double? = null, confidenceLevel: Double = 0.95, probCorrectSelection: Double = 0.95, showAltCIPlot: Boolean = true, showBoxPlot: Boolean = true): ReportNode.Document(source)

Builds a ReportNode.Document containing a complete multiple-comparison analysis of one response across several experiments stored in this database.

This is the cross-experiment counterpart to the single-experiment toReport entry points above. It composes two pieces of existing KSL infrastructure: multipleComparisonAnalyzerFor (which builds a MultipleComparisonAnalyzer from the per-replication values stored for the named experiments) and the multipleComparison report section (which renders alternative statistics, pairwise differences, MCB max/min intervals with confidence-interval plots, and screening results). No statistics are recomputed here and no HTML is hand-built; the document is assembled entirely from the report DSL.

Zero-code path:

val db = KSLDatabase("pharmacy.db")
db.toComparisonReport("SystemTime", listOf("Base", "AddServer", "AddKiosk"))
.showInBrowser()
db.toComparisonReport("SystemTime", listOf("Base", "AddServer"))
.writeMarkdown()

All experiments named in expNames must record responseName with equal replication counts — the underlying MultipleComparisonAnalyzer requires this. An IllegalArgumentException is thrown when an experiment name is not present in the database (propagated from multipleComparisonAnalyzerFor).

Return

the assembled ReportNode.Document

Parameters

responseName

the response, time-weighted, or counter name to compare

expNames

the experiments to compare; each must be present in the database

title

document title; defaults to "Comparison — "

direction

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

indifferenceZone

delta for MCB interval construction; null (default) uses the analyzer's own default indifference zone

confidenceLevel

confidence level for the alternative and pairwise-difference tables; must be in (0, 1); default 0.95

probCorrectSelection

probability of correct selection for screening; default 0.95

showAltCIPlot

when true (default) include the per-alternative confidence-interval plot

showBoxPlot

when true (default) include the per-alternative box plot