kumulant/com.eignex.kumulant.bandit.contextual/KnnContextualSpec

KnnContextualSpec

@Serializable

@SerialName(value = "KnnContextual")

data class KnnContextualSpec(val nbrArms: Int, val k: Int = 5, val maxHistoryPerArm: Int = 1024, val coldStartScore: Double = 1.0, val exploration: Double = 1.0, val distance: String = "squaredL2") : ContextualBanditSpec(source)

Spec for KnnContextualBandit. distance is a named lookup against a small built-in registry; currently "squaredL2" is the only stock entry.

Constructors

KnnContextualSpec

constructor(nbrArms: Int, k: Int = 5, maxHistoryPerArm: Int = 1024, coldStartScore: Double = 1.0, exploration: Double = 1.0, distance: String = "squaredL2")

Properties

coldStartScore

val coldStartScore: Double

Score assigned to arms with no history yet.

distance

val distance: String

Name of the distance function in the factory's registry.

exploration

val exploration: Double

UCB-style exploration scale.

val k: Int

Neighbourhood size used for scoring.

maxHistoryPerArm

val maxHistoryPerArm: Int

Maximum observations retained per arm before FIFO eviction.

nbrArms

val nbrArms: Int

Number of arms in the population.

Functions

materialize

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

Dispatch any ContextualBanditSpec to its concrete bandit.

fun KnnContextualSpec.materialize(random: Random = Random.Default, distanceRegistry: Map<String, (VectorView, VectorView) -> Double> = knnDistanceRegistry): KnnContextualBandit

Build a live KnnContextualBandit from its spec, resolving the distance function via distanceRegistry (defaults to knnDistanceRegistry).

KnnContextualSpec

constructor(nbrArms: Int, k: Int = 5, maxHistoryPerArm: Int = 1024, coldStartScore: Double = 1.0, exploration: Double = 1.0, distance: String = "squaredL2")(source)