kumulant

kumulant/com.eignex.kumulant.bandit.contextual/RegressionContextualSpec

RegressionContextualSpec

@Serializable

@SerialName(value = "RegressionContextual")

data class RegressionContextualSpec(val nbrArms: Int, val regression: LinearRegressionSpec, val posterior: LinearPosterior<*>, val exploration: Double = 1.0, val globalRegression: LinearRegressionSpec? = null) : ContextualBanditSpec(source)

Spec for RegressionContextualBandit with a linear-posterior backbone. The regression variant picks one of the three LinearRegressionResult-typed regressors; the posterior selects the matching scoring rule.

Constructors

RegressionContextualSpec

constructor(nbrArms: Int, regression: LinearRegressionSpec, posterior: LinearPosterior<*>, exploration: Double = 1.0, globalRegression: LinearRegressionSpec? = null)

Properties

val exploration: Double

Per-evaluate exploration scale forwarded to posterior.

globalRegression

val globalRegression: LinearRegressionSpec?

Template for the global pooling regressor; null disables pooling.

val nbrArms: Int

Number of arms in the population.

val posterior: LinearPosterior<*>

Stateless arm scorer applied to each per-arm snapshot.

val regression: LinearRegressionSpec

Per-arm regressor template; cloned for each arm.

Functions

fun ContextualBanditSpec.materialize(random: Random = Random.Default, concurrency: Concurrency = Concurrency.None): Bandit

Dispatch any ContextualBanditSpec to its concrete bandit.

fun RegressionContextualSpec.materialize(random: Random = Random.Default, concurrency: Concurrency = Concurrency.None): RegressionContextualBandit<out LinearRegressionResult>

Build a live RegressionContextualBandit from its spec.

RegressionContextualSpec

constructor(nbrArms: Int, regression: LinearRegressionSpec, posterior: LinearPosterior<*>, exploration: Double = 1.0, globalRegression: LinearRegressionSpec? = null)(source)

exploration

val exploration: Double(source)

Per-evaluate exploration scale forwarded to posterior.

globalRegression

val globalRegression: LinearRegressionSpec?(source)

Template for the global pooling regressor; null disables pooling.

nbrArms

val nbrArms: Int(source)

Number of arms in the population.

posterior

val posterior: LinearPosterior<*>(source)

Stateless arm scorer applied to each per-arm snapshot.

regression

val regression: LinearRegressionSpec(source)

Per-arm regressor template; cloned for each arm.