kumulant

RegressionPosterior

Stateless scorer over a regression snapshot at a query point x. Generalises the "score this arm under the current model and context" loop across linear regressors, trees, and any future regressor type:

  • For linear models the score is bias + x . sample(weights) (Thompson) or bias + alpha * sqrt(xT * Sigma * x) (UCB); see com.eignex.kumulant.stat.regression.glm.LinearPosterior.

  • For tree models the snapshot is routed via findLeaf(x) and the leaf's weighted-variance summary drives the draw; see the TreePosterior family.

Used by com.eignex.kumulant.bandit.contextual.RegressionContextualBandit and by Bayesian- optimisation acquisition functions that need to score candidate points.

Inheritors

Functions

Link copied to clipboard
abstract fun evaluate(snapshot: R, x: VectorView, rng: Random, exploration: Double = 1.0): Double

Score a query point x under the regression snapshot. exploration controls the posterior-variance scale (Thompson) or the UCB width (LinUcb-style); 0.0 collapses to the point estimate.

evaluate

abstract fun evaluate(snapshot: R, x: VectorView, rng: Random, exploration: Double = 1.0): Double(source)

Score a query point x under the regression snapshot. exploration controls the posterior-variance scale (Thompson) or the UCB width (LinUcb-style); 0.0 collapses to the point estimate.