AdmissibleCutPlot

class AdmissibleCutPlot(data: DoubleArray, certificate: AdmissibilityCertificate, variableName: String = "observed value") : BasePlot(source)

Which cut positions a run of equal values rules out.

This is about ties, and only about ties. A partition is defined by intervals of value, so observations that are equal must stay in the same group; a position inside a run of equal values is therefore not a position at all. That is the whole of what AdmissibilityCertificate decides per position — the minimum group size and the minimum distinct count are properties of a whole partition rather than of a single cut, and are not what this figure shows.

On continuous data it has nothing to say, because every observation is distinct and no position is ruled out. Check hasRefusedPositions before drawing it: a figure in which every mark is the same colour tells a reader nothing and invites them to think it is broken. It earns its place on data that has been rounded or recorded to a coarse unit, which is common enough in service-time data to be worth having.

Parameters

data

the sample, which need not be sorted

certificate

the admissibility certificate for that sample

variableName

what the observations measure, for the axis label

Constructors

Link copied to clipboard
constructor(data: DoubleArray, certificate: AdmissibilityCertificate, variableName: String = "observed value")

Properties

Link copied to clipboard

An optional caption. Left empty by default: the plot library does not wrap one, so a long caption runs off the canvas and is clipped. Explanation belongs in the surrounding text.

Link copied to clipboard

Whether any interior position is ruled out at all.

Functions

Link copied to clipboard
open override fun buildPlot(): Plot

Builds a new instance of a Lets-Plot representation of the plot