baybe.recommenders.pure.bayesian.botorch.discrete.recommend_discrete_with_subsets

baybe.recommenders.pure.bayesian.botorch.discrete.recommend_discrete_with_subsets(recommender: BotorchRecommender, subspace_discrete: SubspaceDiscrete, candidates_exp: DataFrame, batch_size: int)[source]

Recommend from a discrete space with subset-generating constraints.

Splits the candidate set into subsets according to subset-generating constraints, runs optimization on each feasible subset, and returns the batch with the highest joint acquisition value. Subsets with fewer candidates than batch_size are skipped.

Parameters:

  • recommender (BotorchRecommender) – The recommender instance.

  • subspace_discrete (SubspaceDiscrete) – The discrete subspace from which to generate recommendations.

  • candidates_exp (DataFrame) – The experimental representation of candidates.

  • batch_size (int) – The size of the recommendation batch.

Return type:

Index

Returns:

The dataframe indices of the recommended points.