Class containing the Model Information describing the formal Spatial (or Temporal) Characteristics of the (set of) random variable(s) under study. More...

#include <Model.hpp>

Inheritance diagram for Model:

Public Member Functions
	Model (const CovContext &ctxt=CovContext())

	Model (int nvar, int ndim=2)

	Model (const Model &m)

Model &	operator= (const Model &m)

virtual	~Model ()

virtual String	toString (const AStringFormat *strfmt=nullptr) const override
	ICloneable interface. More...

int	resetFromDb (const Db *db)

void	setCovList (const ACovAnisoList *covalist)

void	addCov (const CovAniso *cov)

void	addCovFromParam (const ECov &type, double range=0., double sill=1., double param=1., const VectorDouble &ranges=VectorDouble(), const VectorDouble &sills=VectorDouble(), const VectorDouble &angles=VectorDouble(), bool flagRange=true)

void	delCova (int icov)

void	delAllCovas ()

void	setDriftList (const DriftList *driftlist)

void	setDriftIRF (int order=0, int nfex=0)

void	addDrift (const ADrift *drift)

void	setDrifts (const VectorString &driftSymbols)

void	delDrift (int rank)

void	delAllDrifts ()

int	setAnam (const AAnam *anam, const VectorInt &strcnt=VectorInt())

int	unsetAnam ()

bool	isFlagGradient () const

bool	isFlagGradientNumerical () const

bool	isFlagGradientFunctional () const

bool	isFlagLinked () const

CovAniso	extractCova (int icov) const

void	switchToGradient ()

const ACovAnisoList *	getCovAnisoList () const
	TODO : to be removed (encapsulation of ACovAnisoList) More...

const CovAniso *	getCova (unsigned int icov) const

CovAniso *	getCova (unsigned int icov)

int	getCovaNumber (bool skipNugget=false) const

const ECov &	getCovaType (int icov) const

const MatrixSquareSymmetric	getSillValues (int icov) const

double	getSill (int icov, int ivar, int jvar) const

double	getParam (int icov) const

bool	isCovaFiltered (int icov) const

bool	isStationary () const

String	getCovName (int icov) const

int	getGradParamNumber (int icov) const

double	getTotalSill (int ivar=0, int jvar=0) const

MatrixSquareGeneral	getTotalSills () const

double	getBallRadius () const

const AnamHermite *	getAnamHermite () const

double	getMaximumDistance () const

int	getCovaMinIRFOrder () const

bool	hasAnam () const

const AAnam *	getAnam () const

bool	isChangeSupportDefined () const

void	normalize (double sill)

bool	hasNugget () const

VectorInt	getActiveCovList () const

VectorInt	getAllActiveCovList () const

bool	isAllActiveCovList () const

void	setTapeRange (double range)

void	setOptimEnabled (bool flagOptim)

bool	isOptimEnabled () const

double	eval0 (int ivar=0, int jvar=0, const CovCalcMode *mode=nullptr) const

MatrixSquareGeneral	eval0Nvar (const CovCalcMode *mode=nullptr) const

int	isNoStat () const

const ANoStat *	getNoStat () const

ANoStat *	getNoStatModify () const

void	eval0MatInPlace (MatrixSquareGeneral &mat, const CovCalcMode *mode=nullptr) const

double	eval (const SpacePoint &p1, const SpacePoint &p2, int ivar=0, int jvar=0, const CovCalcMode *mode=nullptr) const

MatrixSquareGeneral	evalNvarIpas (double step, const VectorDouble &dir, const CovCalcMode *mode=nullptr) const

MatrixSquareGeneral	evalMat (const SpacePoint &p1, const SpacePoint &p2, const CovCalcMode *mode=nullptr) const

void	evalMatInPlace (const SpacePoint &p1, const SpacePoint &p2, MatrixSquareGeneral &mat, const CovCalcMode *mode=nullptr) const

MatrixSquareGeneral	evalNvarIpasIncr (const VectorDouble &dincr, const CovCalcMode *mode=nullptr) const

VectorDouble	evalIvarNpas (const VectorDouble &vec_step, const VectorDouble &dir=VectorDouble(), int ivar=0, int jvar=0, const CovCalcMode *mode=nullptr) const

double	evalIvarIpas (double step, const VectorDouble &dir=VectorDouble(), int ivar=0, int jvar=0, const CovCalcMode *mode=nullptr) const

double	evalCvv (const VectorDouble &ext, const VectorInt &ndisc, const VectorDouble &angles=VectorDouble(), int ivar=0, int jvar=0, const CovCalcMode *mode=nullptr) const

double	evalCvvShift (const VectorDouble &ext, const VectorInt &ndisc, const VectorDouble &shift, const VectorDouble &angles=VectorDouble(), int ivar=0, int jvar=0, const CovCalcMode *mode=nullptr) const

MatrixSquareGeneral	evalCvvM (const VectorDouble &ext, const VectorInt &ndisc, const VectorDouble &angles=VectorDouble(), const CovCalcMode *mode=nullptr)

double	evalCxv (const SpacePoint &p1, const VectorDouble &ext, const VectorInt &ndisc, const VectorDouble &angles=VectorDouble(), const VectorDouble &x0=VectorDouble(), int ivar=0, int jvar=0, const CovCalcMode *mode=nullptr)

MatrixSquareGeneral	evalCxvM (const SpacePoint &p1, const VectorDouble &ext, const VectorInt &ndisc, const VectorDouble &angles=VectorDouble(), const VectorDouble &x0=VectorDouble(), const CovCalcMode *mode=nullptr)

VectorDouble	evalPointToDb (const SpacePoint &p1, const Db db2, int ivar=0, int jvar=0, bool useSel=true, const VectorInt &nbgh2=VectorInt(), const CovCalcMode mode=nullptr)

VectorDouble	evalPointToDbAsSP (const std::vector< SpacePoint > &p1s, const SpacePoint &p2, int ivar=0, int jvar=0, const CovCalcMode *mode=nullptr) const

double	evalAverageDbToDb (const Db db1, const Db db2, int ivar=0, int jvar=0, double eps=0., int seed=434132, const CovCalcMode *mode=nullptr) const

double	evalAverageIncrToIncr (const VectorVectorDouble &d1, const VectorVectorDouble &d2, int ivar=0, int jvar=0, const CovCalcMode *mode=nullptr) const

double	evalAveragePointToDb (const SpacePoint &p1, const Db db2, int ivar=0, int jvar=0, const CovCalcMode mode=nullptr)

MatrixRectangular	evalCovMatrix (Db db1, Db db2=nullptr, int ivar=0, int jvar=0, const VectorInt &nbgh1=VectorInt(), const VectorInt &nbgh2=VectorInt(), const CovCalcMode *mode=nullptr)

MatrixSparse *	evalCovMatrixSparse (Db db1, Db db2=nullptr, int ivar0=0, int jvar0=0, const VectorInt &nbgh1=VectorInt(), const VectorInt &nbgh2=VectorInt(), const CovCalcMode *mode=nullptr, double eps=EPSILON3)

void	evalMatOptimInPlace (int icas1, int iech1, int icas2, int iech2, MatrixSquareGeneral &mat, const CovCalcMode *mode=nullptr) const

VectorVectorDouble	evalCovMatrixOptim (const Db db1, const Db db2=nullptr, int ivar=0, int jvar=0, const CovCalcMode *mode=nullptr)

double	extensionVariance (const Db *db, const VectorDouble &ext, const VectorInt &ndisc, const VectorDouble &angles=VectorDouble(), const VectorDouble &x0=VectorDouble(), int ivar=0, int jvar=0)

double	samplingDensityVariance (const Db *db, const VectorDouble &ext, const VectorInt &ndisc, const VectorDouble &angles=VectorDouble(), const VectorDouble &x0=VectorDouble(), int ivar=0, int jvar=0) const

double	specificVolume (const Db *db, double mean, const VectorDouble &ext, const VectorInt &ndisc, const VectorDouble &angles=VectorDouble(), const VectorDouble &x0=VectorDouble(), int ivar=0, int jvar=0) const

double	coefficientOfVariation (const Db *db, double volume, double mean, const VectorDouble &ext, const VectorInt &ndisc, const VectorDouble &angles=VectorDouble(), const VectorDouble &x0=VectorDouble(), int ivar=0, int jvar=0) const

double	specificVolumeFromCoV (Db *db, double cov, double mean, const VectorDouble &ext, const VectorInt &ndisc, const VectorDouble &angles=VectorDouble(), const VectorDouble &x0=VectorDouble(), int ivar=0, int jvar=0) const

void	evalZAndGradients (const SpacePoint &p1, const SpacePoint &p2, double &covVal, VectorDouble &covGp, VectorDouble &covGG, const CovCalcMode *mode=nullptr, bool flagGrad=false) const

void	evalZAndGradients (const VectorDouble &vec, double &covVal, VectorDouble &covGp, VectorDouble &covGG, const CovCalcMode *mode=nullptr, bool flagGrad=false) const

double	evalCov (const VectorDouble &incr, int icov=0, const ECalcMember &member=ECalcMember::fromKey("LHS")) const

void	setSill (int icov, int ivar, int jvar, double value)

void	setCovaFiltered (int icov, bool filtered)

void	updateCovByPoints (int icas1, int iech1, int icas2, int iech2)

void	updateCovByMesh (int imesh)

void	setActiveFactor (int iclass)

int	getActiveFactor () const

int	getAnamNClass () const

const DriftList *	getDriftList () const
	TODO : to be removed (encapsulation of DriftList) More...

const ADrift *	getDrift (int il) const

ADrift *	getDrift (int il)

int	getDriftNumber () const

int	getExternalDriftNumber () const

int	getRankFext (int il) const

const VectorDouble &	getDriftCLs () const

double	getDriftCL (int ivar, int il, int ib) const

int	getDriftEquationNumber () const

bool	isDriftFiltered (unsigned int il) const

bool	isDriftDefined (const VectorInt &powers, int rank_fex=0) const

bool	isDriftDifferentDefined (const VectorInt &powers, int rank_fex=-1) const

int	getDriftMaxIRFOrder (void) const

void	resetDriftCoef ()

void	setDriftCoef (int ivar, int il, int ib, double coeff)

void	setDriftFiltered (int il, bool filtered)

VectorDouble	getDriftByColumn (const Db *db, int ib, bool useSel=true)

VectorVectorDouble	getDrifts (const Db *db, bool useSel=true)

void	setBetaHat (const VectorDouble &betaHat)

double	evalDrift (const Db *db, int iech, int il, const ECalcMember &member=ECalcMember::fromKey("LHS")) const

double	evalDriftValue (int ivar, int ib, const VectorDouble &drftab) const

VectorDouble	evalDriftVec (const Db *db, int iech, const ECalcMember &member=ECalcMember::fromKey("LHS")) const

void	evalDriftVecInPlace (const Db *db, int iech, const ECalcMember &member, VectorDouble &drftab) const

double	evalDriftVarCoef (const Db *db, int iech, int ivar, const VectorDouble &coeffs) const

VectorDouble	evalDriftVarCoefVec (const Db *db, const VectorDouble &coeffs, int ivar=0, bool useSel=false) const

MatrixRectangular	evalDriftMat (const Db *db, const ECalcMember &member=ECalcMember::fromKey("LHS")) const

const CovContext &	getContext () const
	TODO : to be removed (encapsulation of Context) More...

const VectorDouble &	getMeans () const

double	getMean (int ivar) const

const VectorDouble &	getCovar0s () const

double	getCovar0 (int ivar, int jvar) const

double	getField () const

int	getDimensionNumber () const

void	setMeans (const VectorDouble &mean)

void	setMean (double mean, int ivar=0)

void	setCovar0s (const VectorDouble &covar0)

void	setCovar0 (int ivar, int jvar, double covar0)

void	setField (double field)

int	addNoStat (const ANoStat *anostat)
	Shortcut for Non-stationary. More...

int	getNoStatElemNumber () const

int	addNoStatElem (int igrf, int icov, const EConsElem &type, int iv1, int iv2)

int	addNoStatElems (const VectorString &codes)

CovParamId	getCovParamId (int ipar) const

bool	isNostatParamDefined (const EConsElem &type0)

const EModelProperty &	getCovMode () const

Model *	duplicate () const

Model *	createReduce (const VectorInt &validVars) const

int	getVariableNumber () const

int	hasExternalCov () const

MatrixSquareSymmetric	covMatrixMS (Db db1, const CovCalcMode mode=nullptr)

MatrixRectangular	covMatrixM (Db db1, Db db2=nullptr, int ivar=-1, int jvar=-1, const CovCalcMode *mode=nullptr)

VectorDouble	covMatrixV (Db db1, Db db2=nullptr, int ivar=0, int jvar=0, const CovCalcMode *mode=nullptr)

VectorDouble	sampleUnitary (const VectorDouble &hh, int ivar=0, int jvar=0, VectorDouble codir=VectorDouble(), const CovCalcMode *mode=nullptr)

VectorDouble	envelop (const VectorDouble &hh, int ivar=0, int jvar=0, int isign=1, VectorDouble codir=VectorDouble(), const CovCalcMode *mode=nullptr)

int	fitFromCovIndices (Vario *vario, const VectorECov &types=ECov::fromKeys({"EXPONENTIAL"}), const Constraints &constraints=Constraints(), Option_VarioFit optvar=Option_VarioFit(), Option_AutoFit mauto=Option_AutoFit(), bool verbose=false)

int	fit (Vario *vario, const VectorECov &types=ECov::fromKeys({"SPHERICAL"}), const Constraints &constraints=Constraints(), Option_VarioFit optvar=Option_VarioFit(), Option_AutoFit mauto=Option_AutoFit(), bool verbose=false)

int	fitFromVMap (DbGrid *dbmap, const VectorECov &types=ECov::fromKeys({"SPHERICAL"}), const Constraints &constraints=Constraints(), Option_VarioFit optvar=Option_VarioFit(), Option_AutoFit mauto=Option_AutoFit(), bool verbose=false)

int	buildVmapOnDbGrid (DbGrid *dbgrid, const NamingConvention &namconv=NamingConvention("VMAP"))

int	stabilize (double percent, bool verbose=false)

int	standardize (bool verbose=false)

double	gofToVario (const Vario *vario, bool verbose=true)

void	gofDisplay (double gof, bool byValue=true, const VectorDouble &thresholds={2., 5., 10., 100})

VectorECov	initCovList (const VectorInt &covranks)

bool	isValid () const

VectorDouble	sample (const VectorDouble &h, const VectorDouble &codir=VectorDouble(), int ivar=0, int jvar=0, const CovCalcMode mode=nullptr, const CovInternal covint=nullptr)

double	evaluateOneIncr (double hh, const VectorDouble &codir=VectorDouble(), int ivar=0, int jvar=0, const CovCalcMode *mode=nullptr)

void	evaluateMatInPlace (const CovInternal covint, const VectorDouble &d1, MatrixSquareGeneral &covtab, bool flag_init=false, double weight=1., const CovCalcMode mode=nullptr)

double	evaluateOneGeneric (const CovInternal covint, const VectorDouble &d1=VectorDouble(), double weight=1., const CovCalcMode mode=nullptr)

VectorDouble	evaluateFromDb (Db db, int ivar=0, int jvar=0, const CovCalcMode mode=nullptr)

double	calculateStdev (Db db1, int iech1, Db db2, int iech2, bool verbose=false, double factor=1., const CovCalcMode *mode=nullptr)

void	nostatUpdate (CovInternal *covint)

double	computeLogLikelihood (Db *db, bool verbose=false)

Public Member Functions inherited from AStringable
	AStringable ()

	AStringable (const AStringable &r)

AStringable &	operator= (const AStringable &r)

virtual	~AStringable ()

virtual void	display (const AStringFormat *strfmt=nullptr) const final

virtual void	display (int level) const final

Public Member Functions inherited from ASerializable
	ASerializable ()

	ASerializable (const ASerializable &r)

ASerializable &	operator= (const ASerializable &r)

virtual	~ASerializable ()

bool	deserialize (std::istream &is, bool verbose=true)

bool	serialize (std::ostream &os, bool verbose=true) const

bool	dumpToNF (const String &neutralFilename, bool verbose=false) const

Public Member Functions inherited from ICloneable
	ICloneable ()

virtual	~ICloneable ()

virtual ICloneable *	clone () const =0

Static Public Member Functions
static Model *	create (const CovContext &ctxt=CovContext())

static Model *	createFromEnvironment (int nvar, int ndim=2)

static Model *	createNugget (int nvar, int ndim=2, double sill=1.)

static Model *	createFromParam (const ECov &type=ECov::fromKey("NUGGET"), double range=1., double sill=1., double param=1., const VectorDouble &ranges=VectorDouble(), const VectorDouble &sills=VectorDouble(), const VectorDouble &angles=VectorDouble(), const ASpace *space=nullptr, bool flagRange=true)

static Model *	createFromDb (const Db *db)

static Model *	createFromNF (const String &neutralFilename, bool verbose=true)

Static Public Member Functions inherited from ASerializable
static String	buildFileName (int status, const String &filename, bool ensureDirExist=false)

static String	getHomeDirectory (const String &sub="")

static String	getWorkingDirectory ()

static String	getTestData (const String &subdir, const String &filename)

static String	getFileIdentity (const String &filename, bool verbose=false)

static void	setContainerName (bool useDefault, const String &containerName="", bool verbose=false)

static void	unsetContainerName ()

static void	setPrefixName (const String &prefixName)

static void	unsetPrefixName ()

static const String &	getContainerName ()

static const String &	getPrefixName ()

static bool	createDirectory (const String &dir)

static String	getExecDirectory ()

static String	getDirectory (const String &path)

Detailed Description

Class containing the Model Information describing the formal Spatial (or Temporal) Characteristics of the (set of) random variable(s) under study.

The Model is essentially a container with two main contents:

the covariance part: see ACov.hpp for more information
the drift part: see DriftList.hpp for more information

The additional member CovContext only serves in carrying the following information:

the number of variables: if more than 1, the Model becomes multivariate
the field extension: this information is needed to get a stationary version to any covariance
the experimental mean vector and the variance-covariance matrix (used to calibrate the Model)

Constructor & Destructor Documentation

Model::Model ( const CovContext & ctxt = CovContext() )

Model::Model	(	int	nvar,
		int	ndim = `2`
	)

Model::Model ( const Model & m )

Model::~Model ( )

virtual

Member Function Documentation

void Model::addCov ( const CovAniso * cov )

void Model::addCovFromParam	(	const ECov &	type,
		double	range = `0.`,
		double	sill = `1.`,
		double	param = `1.`,
		const VectorDouble &	ranges = `VectorDouble()`,
		const VectorDouble &	sills = `VectorDouble()`,
		const VectorDouble &	angles = `VectorDouble()`,
		bool	flagRange = `true`
	)

void Model::addDrift ( const ADrift * drift )

int Model::addNoStat ( const ANoStat * anostat )

Shortcut for Non-stationary.

Define Non-stationary parameters

Parameters

anostat ANoStat pointer will be duplicated

Returns: Error return code

int Model::addNoStatElem	(	int	igrf,
		int	icov,
		const EConsElem &	type,
		int	iv1,
		int	iv2
	)

int Model::addNoStatElems ( const VectorString & codes )

int Model::buildVmapOnDbGrid	(	DbGrid *	dbgrid,
		const NamingConvention &	namconv = `NamingConvention("VMAP")`
	)

Calculate the variogram map from a Model (presented as Variogram, not Covariance)

Returns: Error return code

Parameters

[in]	dbgrid	Grid structure
[in]	namconv	Naming convention

double Model::calculateStdev	(	Db *	db1,
		int	iech1,
		Db *	db2,
		int	iech2,
		bool	verbose = `false`,
		double	factor = `1.`,
		const CovCalcMode *	mode = `nullptr`
	)

Returns the standard deviation at a given increment for a given model between two samples of two Dbs

Parameters

[in]	db1	First Db
[in]	iech1	Rank in the first Db
[in]	db2	Second Db
[in]	iech2	Rank in the second Db
[in]	verbose	Verbose flag
[in]	factor	Multiplicative factor for standard deviation
[in]	mode	CovCalcMode structure

double Model::coefficientOfVariation	(	const Db *	db,
		double	volume,
		double	mean,
		const VectorDouble &	ext,
		const VectorInt &	ndisc,
		const VectorDouble &	angles = `VectorDouble()`,
		const VectorDouble &	x0 = `VectorDouble()`,
		int	ivar = `0`,
		int	jvar = `0`
	)		const

inline

double Model::computeLogLikelihood	(	Db *	db,
		bool	verbose = `false`
	)

Compute the log-likelihood (traditional method)

Parameters

db	Db structure where variable are loaded from
verbose	Verbose flag

Remarks: The calculation considers all the active samples.; It can work in multivariate case with or without drift conditions (linked or not); The algorithm is stopped (with a message) in the heterotopic case // TODO; improve for heterotopic case

MatrixRectangular Model::covMatrixM	(	Db *	db1,
		Db *	db2 = `nullptr`,
		int	ivar = `-1`,
		int	jvar = `-1`,
		const CovCalcMode *	mode = `nullptr`
	)

MatrixSquareSymmetric Model::covMatrixMS	(	Db *	db1,
		const CovCalcMode *	mode = `nullptr`
	)

Returns the covariance matrix for all samples and all variables By construction, this matrix is square and symmetrical

Parameters

db1	Pointer to the Db structure
mode	CovCalcMode description

VectorDouble Model::covMatrixV	(	Db *	db1,
		Db *	db2 = `nullptr`,
		int	ivar = `0`,
		int	jvar = `0`,
		const CovCalcMode *	mode = `nullptr`
	)

Calculate the covariance matrix between active samples of Db1 and active samples of Db2.

Parameters

db1	First Data Base
db2	Second Data Base (if not provided, the first Db is provided instead)
ivar	Rank of the first variable (-1 for all variables)
jvar	Rank of the second variable (-1 for all variables)
mode	CovCalcMode structure

Remarks: The returned argument must have been dimensioned beforehand to (nvar * nechA)^2 where:; -nvar stands for the number of (active) variables; -nechA stands for the number of active samples

Model * Model::create ( const CovContext & ctxt = CovContext() )

static

Model * Model::createFromDb ( const Db * db )

static

Model * Model::createFromEnvironment	(	int	nvar,
		int	ndim = `2`
	)

static

Model * Model::createFromNF	(	const String &	neutralFilename,
		bool	verbose = `true`
	)

static

Model * Model::createFromParam	(	const ECov &	type = `ECov::fromKey("NUGGET")`,
		double	range = `1.`,
		double	sill = `1.`,
		double	param = `1.`,
		const VectorDouble &	ranges = `VectorDouble()`,
		const VectorDouble &	sills = `VectorDouble()`,
		const VectorDouble &	angles = `VectorDouble()`,
		const ASpace *	space = `nullptr`,
		bool	flagRange = `true`
	)

static

Model * Model::createNugget	(	int	nvar,
		int	ndim = `2`,
		double	sill = `1.`
	)

static

Model * Model::createReduce ( const VectorInt & validVars ) const

void Model::delAllCovas ( )

void Model::delAllDrifts ( )

void Model::delCova ( int icov )

void Model::delDrift ( int rank )

Model * Model::duplicate ( ) const

VectorDouble Model::envelop	(	const VectorDouble &	hh,
		int	ivar = `0`,
		int	jvar = `0`,
		int	isign = `1`,
		VectorDouble	codir = `VectorDouble()`,
		const CovCalcMode *	mode = `nullptr`
	)

double Model::eval	(	const SpacePoint &	p1,
		const SpacePoint &	p2,
		int	ivar = `0`,
		int	jvar = `0`,
		const CovCalcMode *	mode = `nullptr`
	)		const

inline

double Model::eval0	(	int	ivar = `0`,
		int	jvar = `0`,
		const CovCalcMode *	mode = `nullptr`
	)		const

inline

void Model::eval0MatInPlace	(	MatrixSquareGeneral &	mat,
		const CovCalcMode *	mode = `nullptr`
	)		const

inline

Calculate the Matrix of covariance for zero distance

Parameters

mat	Covariance matrix (Dimension: nvar * nvar)
mode	Calculation Options

Remarks: : Matrix 'mat' should be dimensioned and initialized beforehand

MatrixSquareGeneral Model::eval0Nvar ( const CovCalcMode * mode = nullptr ) const

inline

double Model::evalAverageDbToDb	(	const Db *	db1,
		const Db *	db2,
		int	ivar = `0`,
		int	jvar = `0`,
		double	eps = `0.`,
		int	seed = `434132`,
		const CovCalcMode *	mode = `nullptr`
	)		const

inline

double Model::evalAverageIncrToIncr	(	const VectorVectorDouble &	d1,
		const VectorVectorDouble &	d2,
		int	ivar = `0`,
		int	jvar = `0`,
		const CovCalcMode *	mode = `nullptr`
	)		const

inline

double Model::evalAveragePointToDb	(	const SpacePoint &	p1,
		const Db *	db2,
		int	ivar = `0`,
		int	jvar = `0`,
		const CovCalcMode *	mode = `nullptr`
	)

inline

double Model::evalCov	(	const VectorDouble &	incr,
		int	icov = `0`,
		const ECalcMember &	member = `ECalcMember::fromKey("LHS")`
	)		const

MatrixRectangular Model::evalCovMatrix	(	Db *	db1,
		Db *	db2 = `nullptr`,
		int	ivar = `0`,
		int	jvar = `0`,
		const VectorInt &	nbgh1 = `VectorInt()`,
		const VectorInt &	nbgh2 = `VectorInt()`,
		const CovCalcMode *	mode = `nullptr`
	)

inline

VectorVectorDouble Model::evalCovMatrixOptim	(	const Db *	db1,
		const Db *	db2 = `nullptr`,
		int	ivar = `0`,
		int	jvar = `0`,
		const CovCalcMode *	mode = `nullptr`
	)

MatrixSparse* Model::evalCovMatrixSparse	(	Db *	db1,
		Db *	db2 = `nullptr`,
		int	ivar0 = `0`,
		int	jvar0 = `0`,
		const VectorInt &	nbgh1 = `VectorInt()`,
		const VectorInt &	nbgh2 = `VectorInt()`,
		const CovCalcMode *	mode = `nullptr`,
		double	eps = `EPSILON3`
	)

inline

double Model::evalCvv	(	const VectorDouble &	ext,
		const VectorInt &	ndisc,
		const VectorDouble &	angles = `VectorDouble()`,
		int	ivar = `0`,
		int	jvar = `0`,
		const CovCalcMode *	mode = `nullptr`
	)		const

inline

MatrixSquareGeneral Model::evalCvvM	(	const VectorDouble &	ext,
		const VectorInt &	ndisc,
		const VectorDouble &	angles = `VectorDouble()`,
		const CovCalcMode *	mode = `nullptr`
	)

inline

double Model::evalCvvShift	(	const VectorDouble &	ext,
		const VectorInt &	ndisc,
		const VectorDouble &	shift,
		const VectorDouble &	angles = `VectorDouble()`,
		int	ivar = `0`,
		int	jvar = `0`,
		const CovCalcMode *	mode = `nullptr`
	)		const

inline

double Model::evalCxv	(	const SpacePoint &	p1,
		const VectorDouble &	ext,
		const VectorInt &	ndisc,
		const VectorDouble &	angles = `VectorDouble()`,
		const VectorDouble &	x0 = `VectorDouble()`,
		int	ivar = `0`,
		int	jvar = `0`,
		const CovCalcMode *	mode = `nullptr`
	)

inline

MatrixSquareGeneral Model::evalCxvM	(	const SpacePoint &	p1,
		const VectorDouble &	ext,
		const VectorInt &	ndisc,
		const VectorDouble &	angles = `VectorDouble()`,
		const VectorDouble &	x0 = `VectorDouble()`,
		const CovCalcMode *	mode = `nullptr`
	)

inline

double Model::evalDrift	(	const Db *	db,
		int	iech,
		int	il,
		const ECalcMember &	member = `ECalcMember::fromKey("LHS")`
	)		const

Evaluate a given drift function for a given sample

Parameters

db	Db structure
iech	Rank of the target sample
il	Rank of the drift function
member	Member type (used to check filtering)

Returns

MatrixRectangular Model::evalDriftMat	(	const Db *	db,
		const ECalcMember &	member = `ECalcMember::fromKey("LHS")`
	)		const

double Model::evalDriftValue	(	int	ivar,
		int	ib,
		const VectorDouble &	drftab
	)		const

double Model::evalDriftVarCoef	(	const Db *	db,
		int	iech,
		int	ivar,
		const VectorDouble &	coeffs
	)		const

Evaluate the drift with a given sample and a given variable The value is scaled by 'coeffs'

Parameters

[in]	db	Db structure
[in]	iech	Rank of the sample
[in]	ivar	Rank of the variable
[in]	coeffs	Vector of coefficients

VectorDouble Model::evalDriftVarCoefVec	(	const Db *	db,
		const VectorDouble &	coeffs,
		int	ivar = `0`,
		bool	useSel = `false`
	)		const

A vector of the drift evaluation (for all samples)

Parameters

db	Db structure
coeffs	Vector of drift coefficients
ivar	Variable rank (used for constant drift value)
useSel	When TRUE, only non masked samples are returned

Returns: The vector of values

Remarks: When no drift is defined, a vector is returned filled with the variable mean

VectorDouble Model::evalDriftVec	(	const Db *	db,
		int	iech,
		const ECalcMember &	member = `ECalcMember::fromKey("LHS")`
	)		const

void Model::evalDriftVecInPlace	(	const Db *	db,
		int	iech,
		const ECalcMember &	member,
		VectorDouble &	drftab
	)		const

double Model::evalIvarIpas	(	double	step,
		const VectorDouble &	dir = `VectorDouble()`,
		int	ivar = `0`,
		int	jvar = `0`,
		const CovCalcMode *	mode = `nullptr`
	)		const

inline

VectorDouble Model::evalIvarNpas	(	const VectorDouble &	vec_step,
		const VectorDouble &	dir = `VectorDouble()`,
		int	ivar = `0`,
		int	jvar = `0`,
		const CovCalcMode *	mode = `nullptr`
	)		const

inline

MatrixSquareGeneral Model::evalMat	(	const SpacePoint &	p1,
		const SpacePoint &	p2,
		const CovCalcMode *	mode = `nullptr`
	)		const

inline

void Model::evalMatInPlace	(	const SpacePoint &	p1,
		const SpacePoint &	p2,
		MatrixSquareGeneral &	mat,
		const CovCalcMode *	mode = `nullptr`
	)		const

inline

Calculate the Matrix of covariance between two space points

Parameters

p1	Reference of the first space point
p2	Reference of the second space point
mat	Covariance matrix (Dimension: nvar * nvar)
mode	Calculation Options

Remarks: : Matrix 'mat' should be dimensioned and initialized beforehand

void Model::evalMatOptimInPlace	(	int	icas1,
		int	iech1,
		int	icas2,
		int	iech2,
		MatrixSquareGeneral &	mat,
		const CovCalcMode *	mode = `nullptr`
	)		const

inline

Calculate the Matrix of covariance between two elements of two Dbs (defined beforehand)

Parameters

icas1	Origin of the Db containing the first point
iech1	Rank of the first point
icas2	Origin of the Db containing the second point
iech2	Rank of the second point
mat	Covariance matrix (Dimension: nvar * nvar)
mode	Calculation Options

Remarks: : Matrix 'mat' should be dimensioned and initialized beforehand

MatrixSquareGeneral Model::evalNvarIpas	(	double	step,
		const VectorDouble &	dir,
		const CovCalcMode *	mode = `nullptr`
	)		const

inline

MatrixSquareGeneral Model::evalNvarIpasIncr	(	const VectorDouble &	dincr,
		const CovCalcMode *	mode = `nullptr`
	)		const

inline

VectorDouble Model::evalPointToDb	(	const SpacePoint &	p1,
		const Db *	db2,
		int	ivar = `0`,
		int	jvar = `0`,
		bool	useSel = `true`,
		const VectorInt &	nbgh2 = `VectorInt()`,
		const CovCalcMode *	mode = `nullptr`
	)

inline

VectorDouble Model::evalPointToDbAsSP	(	const std::vector< SpacePoint > &	p1s,
		const SpacePoint &	p2,
		int	ivar = `0`,
		int	jvar = `0`,
		const CovCalcMode *	mode = `nullptr`
	)		const

inline

VectorDouble Model::evaluateFromDb	(	Db *	db,
		int	ivar = `0`,
		int	jvar = `0`,
		const CovCalcMode *	mode = `nullptr`
	)

Evaluate the model on a Db

Parameters

[in]	db	Db structure
[in]	ivar	Rank of the first variable
[in]	jvar	Rank of the second variable
[in]	mode	CovCalcMode structure

void Model::evaluateMatInPlace	(	const CovInternal *	covint,
		const VectorDouble &	d1,
		MatrixSquareGeneral &	covtab,
		bool	flag_init = `false`,
		double	weight = `1.`,
		const CovCalcMode *	mode = `nullptr`
	)

Returns the covariances for an increment This is the generic internal function It can be called for stationary or non-stationary case

Parameters

[in]	covint	Internal structure for non-stationarityAddress for the next term after the drift or NULL (for stationary case)
[in]	mode	CovCalcMode structure
[in]	flag_init	Initialize the array beforehand
[in]	weight	Multiplicative weight
[in]	d1	Distance vector
[out]	covtab	Covariance array

double Model::evaluateOneGeneric	(	const CovInternal *	covint,
		const VectorDouble &	d1 = `VectorDouble()`,
		double	weight = `1.`,
		const CovCalcMode *	mode = `nullptr`
	)

Returns the covariance for an increment This is the generic internal function It can be called for stationary or non-stationary case

Parameters

[in]	covint	Internal structure for non-stationarityAddress for the next term after the drift or NULL (for stationary case)
[in]	mode	CovCalcMode structure
[in]	weight	Multiplicative weight
[in]	d1	Distance vector

double Model::evaluateOneIncr	(	double	hh,
		const VectorDouble &	codir = `VectorDouble()`,
		int	ivar = `0`,
		int	jvar = `0`,
		const CovCalcMode *	mode = `nullptr`
	)

Calculate the value of the model for a set of distances

Returns: The model value

Parameters

[in]	ivar	Rank of the first variable
[in]	jvar	Rank of the second variable
[in]	mode	CovCalcMode structure
[in]	codir	Array giving the direction coefficients (optional)
[in]	hh	Vector of increments

void Model::evalZAndGradients	(	const SpacePoint &	p1,
		const SpacePoint &	p2,
		double &	covVal,
		VectorDouble &	covGp,
		VectorDouble &	covGG,
		const CovCalcMode *	mode = `nullptr`,
		bool	flagGrad = `false`
	)		const

void Model::evalZAndGradients	(	const VectorDouble &	vec,
		double &	covVal,
		VectorDouble &	covGp,
		VectorDouble &	covGG,
		const CovCalcMode *	mode = `nullptr`,
		bool	flagGrad = `false`
	)		const

double Model::extensionVariance	(	const Db *	db,
		const VectorDouble &	ext,
		const VectorInt &	ndisc,
		const VectorDouble &	angles = `VectorDouble()`,
		const VectorDouble &	x0 = `VectorDouble()`,
		int	ivar = `0`,
		int	jvar = `0`
	)

inline

CovAniso Model::extractCova ( int icov ) const

int Model::fit	(	Vario *	vario,
		const VectorECov &	types = `ECov::fromKeys({"SPHERICAL"})`,
		const Constraints &	constraints = `Constraints()`,
		Option_VarioFit	optvar = `Option_VarioFit()`,
		Option_AutoFit	mauto = `Option_AutoFit()`,
		bool	verbose = `false`
	)

Automatic Fitting procedure from an experimental Variogram

Parameters

vario	Experimental variogram to be fitted
types	Vector of ECov (see remarks)
constraints	Set of Constraints
optvar	Set of options
mauto	Special parameters for Automatic fitting procedure (instance of Option_AutoFit), for exemple wmode (type of weighting function)
verbose	Verbose option

Remarks: If no list of specific basic structure is specified, the automatic fitting is performed using a single spherical structure by default.

Returns: 0 if no error, 1 otherwise

TODO : What to do with that ?

int Model::fitFromCovIndices	(	Vario *	vario,
		const VectorECov &	types = `ECov::fromKeys({"EXPONENTIAL"})`,
		const Constraints &	constraints = `Constraints()`,
		Option_VarioFit	optvar = `Option_VarioFit()`,
		Option_AutoFit	mauto = `Option_AutoFit()`,
		bool	verbose = `false`
	)

Automatic Fitting procedure

Parameters

vario	Experimental variogram to be fitted
types	Vector of ECov integer values
constraints	Set of Constraints
optvar	Set of options
mauto	Special parameters for Automatic fitting procedure
verbose	Verbose option

Returns: 0 if no error, 1 otherwise

TODO : What to do with that ?

int Model::fitFromVMap	(	DbGrid *	dbmap,
		const VectorECov &	types = `ECov::fromKeys({"SPHERICAL"})`,
		const Constraints &	constraints = `Constraints()`,
		Option_VarioFit	optvar = `Option_VarioFit()`,
		Option_AutoFit	mauto = `Option_AutoFit()`,
		bool	verbose = `false`
	)

Automatic Fitting procedure from A Variogram Map stored on a DbGrid

Parameters

dbmap	DbGrid containing the Variogram Map
types	Vector of ECov
constraints	Set of Constraints
optvar	Set of options
mauto	Special parameters for Automatic fitting procedure (instance of Option_AutoFit), for exemple wmode (type of weighting function)
verbose	Verbose option

Returns: 0 if no error, 1 otherwise

VectorInt Model::getActiveCovList ( ) const

int Model::getActiveFactor ( ) const

VectorInt Model::getAllActiveCovList ( ) const

const AAnam * Model::getAnam ( ) const

const AnamHermite * Model::getAnamHermite ( ) const

int Model::getAnamNClass ( ) const

double Model::getBallRadius ( ) const

Returns the Ball radius (from the first covariance of _covaList)

Returns: Value of the Ball Radius (if defined, i.e. for Numerical Gradient calculation)

const CovContext& Model::getContext ( ) const

inline

TODO : to be removed (encapsulation of Context)

const CovAniso * Model::getCova ( unsigned int icov ) const

CovAniso * Model::getCova ( unsigned int icov )

int Model::getCovaMinIRFOrder ( ) const

const ACovAnisoList * Model::getCovAnisoList ( ) const

TODO : to be removed (encapsulation of ACovAnisoList)

int Model::getCovaNumber ( bool skipNugget = false ) const

double Model::getCovar0	(	int	ivar,
		int	jvar
	)		const

inline

const VectorDouble& Model::getCovar0s ( ) const

inline

const ECov & Model::getCovaType ( int icov ) const

const EModelProperty & Model::getCovMode ( ) const

String Model::getCovName ( int icov ) const

CovParamId Model::getCovParamId ( int ipar ) const

int Model::getDimensionNumber ( ) const

inline

const ADrift * Model::getDrift ( int il ) const

ADrift * Model::getDrift ( int il )

VectorDouble Model::getDriftByColumn	(	const Db *	db,
		int	ib,
		bool	useSel = `true`
	)

double Model::getDriftCL	(	int	ivar,
		int	il,
		int	ib
	)		const

const VectorDouble & Model::getDriftCLs ( ) const

int Model::getDriftEquationNumber ( ) const

const DriftList * Model::getDriftList ( ) const

TODO : to be removed (encapsulation of DriftList)

int Model::getDriftMaxIRFOrder ( void ) const

int Model::getDriftNumber ( ) const

VectorVectorDouble Model::getDrifts	(	const Db *	db,
		bool	useSel = `true`
	)

int Model::getExternalDriftNumber ( ) const

double Model::getField ( ) const

inline

int Model::getGradParamNumber ( int icov ) const

double Model::getMaximumDistance ( ) const

double Model::getMean ( int ivar ) const

inline

const VectorDouble& Model::getMeans ( ) const

inline

const ANoStat* Model::getNoStat ( ) const

inline

int Model::getNoStatElemNumber ( ) const

ANoStat* Model::getNoStatModify ( ) const

inline

double Model::getParam ( int icov ) const

int Model::getRankFext ( int il ) const

double Model::getSill	(	int	icov,
		int	ivar,
		int	jvar
	)		const

const MatrixSquareSymmetric Model::getSillValues ( int icov ) const

double Model::getTotalSill	(	int	ivar = `0`,
		int	jvar = `0`
	)		const

MatrixSquareGeneral Model::getTotalSills ( ) const

int Model::getVariableNumber ( ) const

inline

void Model::gofDisplay	(	double	gof,
		bool	byValue = `true`,
		const VectorDouble &	thresholds = `{2., 5., 10., 100}`
	)

Printout of statement concerning the Quality of the GOF

Parameters

gof	Value of the Gof
byValue	true: display GOF value; false: print its quality level
thresholds	Vector giving the Quality thresholds

double Model::gofToVario	(	const Vario *	vario,
		bool	verbose = `true`
	)

Evaluate the Goodness-of_fit of the Model on the Experimental Variogram It is expressed as the average departure between Model and Variogram scaled to the variance. As this variance may be poorly calculated (< gmax / 5), it may be replaced by the largest value (gmax) divided by 2 (highly non_stationary cases).

Parameters

vario	Experimental variogram
verbose	Verbose flag

Returns: Value for the Goodness-of_fit (as percentage of the total sill)

bool Model::hasAnam ( ) const

int Model::hasExternalCov ( ) const

bool Model::hasNugget ( ) const

VectorECov Model::initCovList ( const VectorInt & covranks )

bool Model::isAllActiveCovList ( ) const

bool Model::isChangeSupportDefined ( ) const

bool Model::isCovaFiltered ( int icov ) const

bool Model::isDriftDefined	(	const VectorInt &	powers,
		int	rank_fex = `0`
	)		const

bool Model::isDriftDifferentDefined	(	const VectorInt &	powers,
		int	rank_fex = `-1`
	)		const

bool Model::isDriftFiltered ( unsigned int il ) const

bool Model::isFlagGradient ( ) const

bool Model::isFlagGradientFunctional ( ) const

bool Model::isFlagGradientNumerical ( ) const

bool Model::isFlagLinked ( ) const

int Model::isNoStat ( ) const

inline

bool Model::isNostatParamDefined ( const EConsElem & type0 )

Check if the non-stationary Model has a given non-stationary parameter

Returns: 1 if the given non-stationary parameter is defined; 0 otherwise

Parameters

[in] type0 Requested type (EConsElem)

bool Model::isOptimEnabled ( ) const

inline

bool Model::isStationary ( ) const

bool Model::isValid ( ) const

void Model::normalize ( double sill )

void Model::nostatUpdate ( CovInternal * covint )

Update the Model in the case of Non-stationary parameters This requires the knowledge of the two end-points

Parameters

[in] covint Internal structure for non-stationarity or NULL (for stationary case)

Model & Model::operator= ( const Model & m )

void Model::resetDriftCoef ( )

int Model::resetFromDb ( const Db * db )

VectorDouble Model::sample	(	const VectorDouble &	h,
		const VectorDouble &	codir = `VectorDouble()`,
		int	ivar = `0`,
		int	jvar = `0`,
		const CovCalcMode *	mode = `nullptr`,
		const CovInternal *	covint = `nullptr`
	)

Calculate the value of the model for a set of distances

Returns: Array containing the model values

Parameters

[in]	ivar	Rank of the first variable
[in]	jvar	Rank of the second variable
[in]	codir	Array giving the direction coefficients (optional)
[in]	h	Vector of increments
[in]	mode	CovCalcMode structure
[in]	covint	Non-stationary parameters

VectorDouble Model::sampleUnitary	(	const VectorDouble &	hh,
		int	ivar = `0`,
		int	jvar = `0`,
		VectorDouble	codir = `VectorDouble()`,
		const CovCalcMode *	mode = `nullptr`
	)

Returns the value of the normalized covariance (by the variance/covariance value) for a given pair of variables

Parameters

hh	Vector of distances
ivar	Rank of the first variable
jvar	Rank of the second variable
codir	Direction coefficients
mode	CovCalcMode structure

Returns

double Model::samplingDensityVariance	(	const Db *	db,
		const VectorDouble &	ext,
		const VectorInt &	ndisc,
		const VectorDouble &	angles = `VectorDouble()`,
		const VectorDouble &	x0 = `VectorDouble()`,
		int	ivar = `0`,
		int	jvar = `0`
	)		const

inline

void Model::setActiveFactor ( int iclass )

int Model::setAnam	(	const AAnam *	anam,
		const VectorInt &	strcnt = `VectorInt()`
	)

Defining an Anamorphosis information for the Model (in fact, this is added to ACovAnisoList part and transforms it from CovLMC to CovLMCAnamorphosis

Parameters

anam	Pointer to the anamorphosis
strcnt	Array of covariance description used for IR case

Returns

void Model::setBetaHat ( const VectorDouble & betaHat )

void Model::setCovaFiltered	(	int	icov,
		bool	filtered
	)

void Model::setCovar0	(	int	ivar,
		int	jvar,
		double	covar0
	)

void Model::setCovar0s ( const VectorDouble & covar0 )

void Model::setCovList ( const ACovAnisoList * covalist )

Add a list of Covariances. This operation cleans any previously stored covariance

Parameters

covalist List of Covariances to be added

void Model::setDriftCoef	(	int	ivar,
		int	il,
		int	ib,
		double	coeff
	)

void Model::setDriftFiltered	(	int	il,
		bool	filtered
	)

void Model::setDriftIRF	(	int	order = `0`,
		int	nfex = `0`
	)

Define the list of drift functions for:

a given degree of the IRF
a given number of external drifts
Parameters

order Order of the IRF

nfex Number of External Drifts

Remarks
This method deletes any pre-existing drift functions and replaces them by the new definition

This replacement is performed accounting for information stored in 'model', such as:
the space dimension
the number of variables

void Model::setDriftList ( const DriftList * driftlist )

Add a list of Drifts. This operation cleans any previously stored drift function

Parameters

driftlist List of Drifts to be added

Remarks: This method deletes any pre-existing drift functions

void Model::setDrifts ( const VectorString & driftSymbols )

void Model::setField ( double field )

void Model::setMean	(	double	mean,
		int	ivar = `0`
	)

void Model::setMeans ( const VectorDouble & mean )

void Model::setOptimEnabled ( bool flagOptim )

inline

void Model::setSill	(	int	icov,
		int	ivar,
		int	jvar,
		double	value
	)

void Model::setTapeRange ( double range )

double Model::specificVolume	(	const Db *	db,
		double	mean,
		const VectorDouble &	ext,
		const VectorInt &	ndisc,
		const VectorDouble &	angles = `VectorDouble()`,
		const VectorDouble &	x0 = `VectorDouble()`,
		int	ivar = `0`,
		int	jvar = `0`
	)		const

inline

double Model::specificVolumeFromCoV	(	Db *	db,
		double	cov,
		double	mean,
		const VectorDouble &	ext,
		const VectorInt &	ndisc,
		const VectorDouble &	angles = `VectorDouble()`,
		const VectorDouble &	x0 = `VectorDouble()`,
		int	ivar = `0`,
		int	jvar = `0`
	)		const

inline

int Model::stabilize	(	double	percent,
		bool	verbose = `false`
	)

Stabilize the model (in the monovariate case)

Returns: Error returned code

Parameters

[in]	percent	Percentage of nugget effect added
[in]	verbose	true for a verbose output

Remarks: If the model only contains GAUSSIAN structures, add; a NUGGET EFFECT structure with a sill equal to a percentage; of the total sill of the GAUSSIAN component(s); This function does not do anything in the multivariate case

int Model::standardize ( bool verbose = false )

Normalize the model

Parameters

[in] verbose true for a verbose output

void Model::switchToGradient ( )

Switch to a Model dedicated to Gradients (transforms it from CovLMC to CovLMGradient)

String Model::toString ( const AStringFormat * strfmt = nullptr ) const

overridevirtual

ICloneable interface.

AStringable Interface

Reimplemented from AStringable.

int Model::unsetAnam ( )

void Model::updateCovByMesh ( int imesh )

void Model::updateCovByPoints	(	int	icas1,
		int	iech1,
		int	icas2,
		int	iech2
	)

The documentation for this class was generated from the following files:

include/Model/Model.hpp
src/Model/Model.cpp

Public Member Functions

Static Public Member Functions

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation