aixd.sampler

This package contains …..

Callbacks

`SamplingEngine`	Initialises the sampling engine.
`AgnosticSamplingEngine`	Represents sampling procedures that do not adapt to objectives and constraints.
`AdaptiveSamplingEngine`	Represents sampling procedures that adapt to objectives and constraints.
`RandomSamplingEngine`	Samples randomly uniform in [0, 1].
`GridamplingEngine`	Returns a grid of size n in each dimension.
`SobolSamplingEngine`	Samples in [0, 1] according to a Sobol sequence.
`LHCSamplingEngine`	Samples in [0, 1] according to a Latin Hypercube sequence.
`BayesOptSamplingEngine`	Samples in [0, 1] according to a Bayesian Optimisation strategy, where the objectives are maximised.

`Operator`	A base class for defining arithmetic and boolean operations on objects that do not yet contain actual data.
`Arithmetic`	A base class for defining arithmetic operations on objects that do not yet contain actual data.
`Boolean`	A base class for defining boolean operations on objects that do not yet contain actual data.
`Constant`	Trivial operator, just returning the stored value.
`Add`	Adding operation using operators.
`Multiply`	Multiply operation using operators.
`Subtract`	Substract operation using operators.
`Divide`	Divide operation using operators.
`LessThan`	Boolean less than operation.
`LessOrEqual`	Boolean less or equal operation.
`GreaterThan`	Boolean greater than operation.
`GreaterOrEqual`	Boolean greater or equal operation.
`Log`	Log operation.
`Exp`	Exponential operation.
`Pow`	Power of base to exponent operation.
`Not`	Not boolean operation.
`And`	And boolean operation.
`Or`	Or boolean operation.
`XOr`	Xor boolean operation.
`Equal`	Equal boolean operation.
`Negative`	Negation operation.
`CastBooleanToConstant`	Type casting operation.

`Reducer`	Perform agregation operations on the output of an operator.
`Sum`	Add up all the values of the operator.
`Mean`	Mean of values of the operator.
`Std`	Standard deviation of values of the operator.
`Var`	Variance of the values of the operator.
`All`	Check if condition is true for all values of the operator.
`Any`	Check if condition is satisfied for any value of the operator.

`sampler_uniform`	Just a sampler to uniformly sample from all design parameters
`sampler_kde`	A KDE sampler that can be fit to some data, and then sample from its distribution
`sampler_quantile`	A sampler that can be fit to some data, and then sample using a quantile strategy.
`sampler_custom`	A sampler to sample from all design parameters, given some data containing the distributions we are intending For each column, aka dataobject, we can just specify a different distribution.
`sampler_conditional_kde`	A KDE sampler fitted to some data, and also link to some conditions.
`sampler_bayesian_kde`	A KDE fitted to some data, and then sampling optimizing for some objective.

Samples values according to certain strategies and by, optionally, optimising objectives and respecting conditions.

`Strategy`	Receives samples from the range [0, 1]^len(features) and maps them to the feature-space according to some strategy.
`UniformStrategy`	Scales and shifts the samples from [0, 1] to [feature.min, feature.max]) for continuous features, or discretises them through binning.
`QuantileStrategy`	Fits a univariate quantile transformer to each of the features.
`KernelDensityStrategy`	Fits a multivariate Kernel Density Estimator to the features.