Decomposition of one or multiple random fields. More...

Public Types
enum	eShapeMethod { EXACT_SVD = 0, RED_SVD, BDC_SVD }
	an enum type for the algorithm computing the shapes More...

Public Member Functions
	check (Structure database)
	Checks the settings on plausibility It throws an exception if there is a problem with the settings. Important: It tries to find out all problems with these settings and then exits the routine. More...

	compute (Structure database)
	Performs the reduction of a random field for multiple sets of samples. More...

	RandomFieldDecompositionFromSamples ()
	default constructor

Public Attributes
eShapeMethod	algorithm
	Defines how the shapes are computed.

bool	compute_amplitudes
	If this is true, then also the amplitudes are computed (a single amplitude per shape per sample).

bool	compute_cumulative_variations
	true, if the cumulative variations are to be computed

bool	compute_variations
	true, if the individual variations are to be computed

bool	cross_correlated
	True if the selected fields are assumed to be cross-correlated, else (default): each field is decomposed individually.

number	desired_variation
	the desired variation (when reducing the model; 0<x<1). Default: 0.90

MultivariateDistributionTypes::enum_type	distribution_type
	The ident of the probability distribution that is assumed to convert data from original space to Gaussian. If empty (default), then it is automatically chosen.

DataObjectContainer	element_data
	the data set to be analyzed

DataObjectIdentMap	element_data_bounds_lower
	lower bounds for ELEMENT_DATA objects

DataObjectIdentMap	element_data_bounds_upper
	upper bounds for ELEMENT_DATA objects

string	field_ident
	If the random fields are cross correlated, they need to have a random field ident being associated.

bool	interpolate_missing_items
	Defines how missing items (eroded elements) are treated: More...

number	max_number
	the maximum number of vectors to be computed. Default: 20

DataObjectContainer	node_data
	the data set to be analyzed

DataObjectIdentMap	node_data_bounds_lower
	lower bounds for NODE_DATA objects

DataObjectIdentMap	node_data_bounds_upper
	upper bounds for NODE_DATA objects

bool	overwrite_existing
	Is true if already existing output data objects will be replaced. In this case, also all depending objects of the already existing one will be deleted.

DataObjectContainer	scalar_data
	the data set to be analyzed

DataObjectIdentMap	scalar_data_bounds_lower
	lower bounds for SCALAR_DATA objects

DataObjectIdentMap	scalar_data_bounds_upper
	upper bounds for SCALAR_DATA objects

Detailed Description

Decomposition of one or multiple random fields.

Member Enumeration Documentation

◆ eShapeMethod

enum eShapeMethod

an enum type for the algorithm computing the shapes

Enumerator
EXACT_SVD	exact full SVD (default up to SoS 7)
RED_SVD	reduced SVD (still needs full memory, but is much faster and less accurate)
BDC_SVD	exact SVD using a divide/conquer iterative algorithm (new default)

Member Function Documentation

◆ check()

check ( Structure database )

Checks the settings on plausibility It throws an exception if there is a problem with the settings. Important: It tries to find out all problems with these settings and then exits the routine.

Parameters

database the SoS structure

◆ compute()

compute ( Structure database )

Performs the reduction of a random field for multiple sets of samples.

Parameters

database The output object where new objects are added to

Member Data Documentation

◆ interpolate_missing_items

bool interpolate_missing_items

Defines how missing items (eroded elements) are treated:

true: use the interpolated value
false: assume the mean value

optiSLang 3D Postprocessing Script API 2025 R1

RandomFieldDecompositionFromSamples Class Reference

Public Types

Public Member Functions

Public Attributes