KSLCore/ksl.utilities.distributions.fitting.diagnostics

Package-level declarations

Types

class ModalityAnalyzer(data: DoubleArray, val numBins: Int = defaultNumBins, streamNum: Int = 0, val streamProvider: RNStreamProviderIfc = KSLRandom.DefaultRNStreamProvider) : RNStreamControlIfc

How many modes the data appears to have, and whether that appearance survives smoothing.

class ModalityAssessment(val apparentModes: Int, val unimodalityTest: ModalityTestResult, val modePersistence: ClosedRange<Double>?, val binSensitivity: Map<Int, Int>, val modeTrace: List<ModeTraceEntry>)

What is known about the shape of the data before anything is fitted to it.

class ModalityTestResult(val maxModes: Int, val criticalBandwidth: Double, val pValue: Double, val numBootstrapSamples: Int)

The outcome of Silverman's test of "at most this many modes".

class ModeTraceEntry(val bandwidth: Double, val numModes: Int, val modeLocations: DoubleArray)

One row of the mode trace: how many modes the estimate has at a bandwidth, and where they are.

The widest span of bandwidths over which the mode count holds at a value, or null when it never does.