Two-Stage Acquisition Policy — DEJAN Codebase Explorer

The acquisition policy selects hypotheses via a two-stage process that samples a Rank Factor Class using freshness weighting to enforce exploration before applying Bayesian selection methods to pick specific claims within that class.

Referenced by

Rank Factor Taxonomy
Rank Factors in Snippet Optimization
Content Optimizer Loop Mechanics
Thompson Sampling in Content Optimizer
Bayesian Acquisition Loop Mechanics
Bayesian Optimization Loop Details
Posterior Win Probability
Rank factor classification
Rank Factor Class Taxonomy
Bayesian Optimization Loop Mechanics
Bayesian Optimization Loop Details
Rank Factor Classification Method
Bayesian Content Optimizer Architecture
Optimizer Acquisition and Sampling
Bayesian Snippet Optimization
Bayesian Acquisition Algorithms
Preventing Model Overfitting