kumulant
com.eignex.kumulant/schema/spec/DiagonalRegression

DiagonalRegression

@Serializable
@SerialName(value = "DiagonalRegression")
data class DiagonalRegression(val featureSize: Int, val priorPrecision: Double = StatDefaults.DIAGONAL_REGRESSION_PRIOR_PRECISION, val learningRate: ScalarExpr = StatDefaults.diagonalLearningRate(), val penalty: Penalty = StatDefaults.regressionPenalty, val link: Link = StatDefaults.regressionLink) : RegressionStatSpec<DiagonalRegressionResult> (source)

Spec for DiagonalRegressionStat: factorised-Gaussian posterior with per-coordinate precision.

Constructors

DiagonalRegression

constructor(featureSize: Int, priorPrecision: Double = StatDefaults.DIAGONAL_REGRESSION_PRIOR_PRECISION, learningRate: ScalarExpr = StatDefaults.diagonalLearningRate(), penalty: Penalty = StatDefaults.regressionPenalty, link: Link = StatDefaults.regressionLink)(source)

Properties

featureSize

val featureSize: Int(source)

Number of input features.

learningRate

val learningRate: ScalarExpr(source)

Per-step learning rate.

penalty

val penalty: Penalty(source)

Gradient-step regulariser.

priorPrecision

val priorPrecision: Double(source)

Initial per-coordinate precision (inverse variance) seeded into every weight.