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) orbias + 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 theTreePosteriorfamily.
Used by com.eignex.kumulant.bandit.contextual.RegressionContextualBandit and by Bayesian- optimisation acquisition functions that need to score candidate points.
Inheritors
Functions
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
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.