smac.model.multi_objective_model

Classes

MultiObjectiveModel(models, objectives[, seed])

Wrapper for the surrogate model to predict multiple objectives.

Interfaces

class smac.model.multi_objective_model.MultiObjectiveModel(models, objectives, seed=0)[source]

Bases: AbstractModel

Wrapper for the surrogate model to predict multiple objectives.

Parameters:

  • models (AbstractModel | list[AbstractModel]) – Which model should be used. If it is a list, then it must provide as many models as objectives. If it is a single model only, the model is used for all objectives.

  • objectives (list[str]) – Which objectives should be used.

  • seed (int) –

property models: list[AbstractModel]

The internally used surrogate models.

predict_marginalized(X)[source]

Predicts mean and variance marginalized over all instances.

Warning

The input data must not include any features.

Parameters:

X (np.ndarray [#samples, #hyperparameters]) – Input data points.

Return type:

tuple[ndarray, ndarray]

Returns:

  • means (np.ndarray [#samples, 1]) – The predictive mean.

  • vars (np.ndarray [#samples, 1]) – The predictive variance.

previous

smac.model.gaussian_process.priors.tophat_prior

next

smac.model.random_forest