bayesvalidrox.surrogate_models.exp_designs.ExpDesigns

class bayesvalidrox.surrogate_models.exp_designs.ExpDesigns(input_object, meta_Model_type='pce', sampling_method='random', hdf5_file=None, n_new_samples=1, n_max_samples=None, mod_LOO_threshold=1e-16, tradeoff_scheme=None, n_canddidate=1, explore_method='random', exploit_method='Space-filling', util_func='Space-filling', n_cand_groups=4, n_replication=1, post_snapshot=False, step_snapshot=1, max_a_post=None, adapt_verbose=False, max_func_itr=1, out_dir='')

Bases: InputSpace

This class generates samples from the prescribed marginals for the model parameters using the Input object.

Attributes

input_objectobj

Input object containing the parameter marginals, i.e. name, distribution type and distribution parameters or available raw data.

meta_Model_typestr

Type of the meta_Model_type.

sampling_methodstr

Name of the sampling method for the experimental design. The following sampling method are supported:

• random

• latin_hypercube

• sobol

• halton

• hammersley

• chebyshev(FT)

• grid(FT)

• user

hdf5_filestr

Name of the hdf5 file that contains the experimental design.

n_new_samplesint

Number of (initial) training points.

n_max_samplesint

Number of maximum training points.

mod_LOO_thresholdfloat

The modified leave-one-out cross validation threshold where the sequential design stops.

tradeoff_schemestr

Trade-off scheme to assign weights to the exploration and exploitation scores in the sequential design.

n_canddidateint

Number of candidate training sets to calculate the scores for.

explore_methodstr

Type of the exploration method for the sequential design. The following methods are supported:

• Voronoi

• random

• latin_hypercube

• LOOCV

• dual annealing

exploit_methodstr

Type of the exploitation method for the sequential design. The following methods are supported:

• BayesOptDesign

• BayesActDesign

• VarOptDesign

• alphabetic

• Space-filling

util_funcstr or list

The utility function to be specified for the exploit_method. For the available utility functions see Note section.

n_cand_groupsint

Number of candidate groups. Each group of candidate training sets will be evaulated separately in parallel.

n_replicationint

Number of replications. Only for comparison. The default is 1.

post_snapshotint

Whether to plot the posterior in the sequential design. The default is True.

step_snapshotint

The number of steps to plot the posterior in the sequential design. The default is 1.

max_a_postlist or array

Maximum a posteriori of the posterior distribution, if known. The default is [].

adapt_verbosebool

Whether to plot the model response vs that of metamodel for the new trining point in the sequential design.

Note

The following utiliy functions for the exploitation methods are supported:

#### BayesOptDesign (when data is available) - DKL (Kullback-Leibler Divergence) - DPP (D-Posterior-percision) - APP (A-Posterior-percision)

#### VarBasedOptDesign -> when data is not available - Entropy (Entropy/MMSE/active learning) - EIGF (Expected Improvement for Global fit) - LOOCV (Leave-one-out Cross Validation)

#### alphabetic - D-Opt (D-Optimality) - A-Opt (A-Optimality) - K-Opt (K-Optimality)

__init__(input_object, meta_Model_type='pce', sampling_method='random', hdf5_file=None, n_new_samples=1, n_max_samples=None, mod_LOO_threshold=1e-16, tradeoff_scheme=None, n_canddidate=1, explore_method='random', exploit_method='Space-filling', util_func='Space-filling', n_cand_groups=4, n_replication=1, post_snapshot=False, step_snapshot=1, max_a_post=None, adapt_verbose=False, max_func_itr=1, out_dir='')

Methods

__init__(input_object[, meta_Model_type, ...])
build_polytypes(rosenblatt)	Creates the polynomial types to be passed to univ_basis_vals method of the MetaModel object.
check_valid_inputs()	Check if the given InputObj is valid to use for further calculations: 1) Has some Marginals 2) The Marginals have valid priors 3) All Marginals given as the same type (samples vs dist)
generate_ED(n_samples[, max_pce_deg])	Generates experimental designs (training set) with the given method.
generate_samples(n_samples[, sampling_method])	Generates samples with given sampling method
init_param_space([max_deg])	Initializes parameter space.
pcm_sampler(n_samples, max_deg)	Generates collocation points based on the root of the polynomial degrees.
plot_samples(samples)	Visualizes generated samples over their given distributions.
random_sampler(n_samples[, max_deg])	Samples the given raw data randomly.
read_from_file(out_names)	Reads in the ExpDesign from a provided h5py file and saves the results.
transform(X[, params, method])	Transforms the samples via either a Rosenblatt or an isoprobabilistic transformation.

build_polytypes(rosenblatt)

Creates the polynomial types to be passed to univ_basis_vals method of the MetaModel object.

Parameters

rosenblattbool

Rosenblatt transformation flag.

Returns

orig_space_distobject

A chaospy JDist object or a gaussian_kde object.

poly_typeslist

A list of polynomial types for the parameters.

check_valid_inputs() → None

Check if the given InputObj is valid to use for further calculations: 1) Has some Marginals 2) The Marginals have valid priors 3) All Marginals given as the same type (samples vs dist)

Returns

None

generate_ED(n_samples, max_pce_deg=1)

Generates experimental designs (training set) with the given method.

Parameters

n_samplesint

Number of requested training points.

max_pce_degint, optional

Maximum PCE polynomial degree. The default is 1.

Returns

None

generate_samples(n_samples, sampling_method='random')

Generates samples with given sampling method

Parameters

n_samplesint

Number of requested samples.

sampling_methodstr, optional

Sampling method. The default is ‘random’.

Returns

samples: array of shape (n_samples, n_params)

Generated samples from defined model input object.

init_param_space(max_deg=1)

Initializes parameter space.

Parameters

max_degint, optional

Maximum degree. The default is 1.

Returns

raw_dataarray of shape (n_params, n_samples)

Raw data.

bound_tupleslist of tuples

A list containing lower and upper bounds of parameters.

pcm_sampler(n_samples, max_deg)

Generates collocation points based on the root of the polynomial degrees.

Parameters

n_samplesint

Number of requested samples.

max_degint

Maximum degree defined by user. Will also be used to run init_param_space if that has not been done beforehand.

Returns

opt_col_points: array of shape (n_samples, n_params)

Collocation points.

plot_samples(samples)

Visualizes generated samples over their given distributions.

Parameters

samplesarray

The samples to visualize.

Returns

None.

random_sampler(n_samples, max_deg=1)

Samples the given raw data randomly.

Parameters

n_samplesint

Number of requested samples.

max_degint, optional

Maximum degree. The default is 1. This will be used to run init_param_space, if it has not been done until now.

Returns

samples: array of shape (n_samples, n_params)

The sampling locations in the input space.

read_from_file(out_names)

Reads in the ExpDesign from a provided h5py file and saves the results.

Parameters

out_nameslist of strings

The keys that are in the outputs (y) saved in the provided file.

Returns

None.

transform(X, params=None, method=None)

Transforms the samples via either a Rosenblatt or an isoprobabilistic transformation.

Parameters

Xarray of shape (n_samples,n_params)

Samples to be transformed.

paramslist

Parameters for laguerre/gamma-type distribution.

methodstring

If transformation method is ‘user’ transform X, else just pass X.

Returns

tr_X: array of shape (n_samples,n_params)

Transformed samples.