QuantileStrategy

class aixd.sampler.strategies.QuantileStrategy(features: DataObject | List[DataObject], n_quantiles: float = 1000, engine: SamplingEngine | str = 'random', **kwargs)

Bases: Strategy

Fits a univariate quantile transformer to each of the features. Allows to sample according to some prior distribution provided by the user through the “fit” function. https://scikit-learn.org/stable/modules/generated/sklearn.preprocessing.QuantileTransformer.html

Parameters:

• features (Union[DataObject, List[DataObject]]) – The features refer to the data objects, which contain the information about the domain of the features.

• n_quantiles (float, optional, default=1000) – Number of quantiles to be computed. It corresponds to the number of landmarks used to discretize the cumulative distribution.

• engine (Union[SamplingEngine, str], optional, default=”random”) – In this case the default engine is “random”, as it is the only one that makes sense to use with this strategy.

Methods

fit	Fit the strategy to the data, also using some vectors of objectives and validity values.
reset_states	Reset the state of the engine.
sample

Inherited Methods

update

Updates the state of the sampling methods by providing the performance of the generated samples in the form of objectives and validity.