Prototypes
LALInferenceKDE *	LALInferenceNewKDE (REAL8 pts, INT4 npts, INT4 dim, INT4 mask)
	Allocate, fill, and tune a Gaussian kernel density estimate from an array of samples. More...

LALInferenceKDE *	LALInferenceNewKDEfromMat (gsl_matrix data, INT4 mask)
	Allocate, fill, and tune a Gaussian kernel density estimate from a matrix of points. More...

LALInferenceKDE *	LALInferenceInitKDE (INT4 npts, INT4 dim)
	Construct an empty KDE structure. More...

void	LALInferenceDestroyKDE (LALInferenceKDE *kde)
	Free an allocated KDE structure. More...

void	LALInferenceSetKDEBandwidth (LALInferenceKDE *kde)
	Calculate the bandwidth and normalization factor for a KDE. More...

REAL8	LALInferenceKDEEvaluatePoint (LALInferenceKDE kde, REAL8 point)
	Evaluate the (log) PDF from a KDE at a single point. More...

REAL8 *	LALInferenceDrawKDESample (LALInferenceKDE kde, gsl_rng rng)
	Draw a sample from a kernel density estimate. More...

INT4	LALInferenceCholeskyDecompose (gsl_matrix *mat)
	Compute the Cholesky decomposition of a matrix. More...

void	LALInferenceComputeCovariance (gsl_matrix cov, gsl_matrix pts)
	Calculate the covariance matrix of a data set. More...

void	LALInferenceComputeMean (gsl_vector mean, gsl_matrix points)
	Calculate the mean of a data set. More...

REAL8	LALInferenceMatrixDet (gsl_matrix *mat)
	Calculate the determinant of a matrix. More...

REAL8	log_add_exps (REAL8 *vals, INT4 size)
	Determine the log of the sum of an array of exponentials. More...

Data Structures
struct	LALInferenceKDE
	Structure containing the Guassian kernel density of a set of samples. More...

Function Documentation

◆ LALInferenceNewKDE()

LALInferenceKDE * LALInferenceNewKDE	(	REAL8 *	pts,
		INT4	npts,
		INT4	dim,
		INT4 *	mask
	)

Allocate, fill, and tune a Gaussian kernel density estimate from an array of samples.

Given a set of points, the distribution from which the points were sampled from is estimated by a kernel density estimate (KDE), using a Gaussian kernal. This routine initializes a new KDE and sets the bandwidth according to Scott's rule. From this it is possible to both evaluate the estimate of the probability density function at an arbitrary location, and generate more samples from the estimated distribution. A mask can be provided to select a subset of the provided data, otherwise all data is used.

Parameters

[in]	pts	Array containing the points to estimate the distribution of.
[in]	npts	The total number of points contained in pts.
[in]	dim	The number of dimensions of the points contained in pts.
[in]	mask	An optional npts long 0/1 array to mask points in pts.

Returns: A LALInferenceKDE structure containing the KDE of the distribution.

Definition at line 63 of file LALInferenceKDE.c.

◆ LALInferenceNewKDEfromMat()

LALInferenceKDE * LALInferenceNewKDEfromMat	(	gsl_matrix *	data,
		INT4 *	mask
	)

Allocate, fill, and tune a Gaussian kernel density estimate from a matrix of points.

Estimate the underlying distribution of samples in a GSL matrix using a Gaussian KDE.

Parameters

[in]	data	GSL matrix with rows of samples from a target distribution.
[in]	mask	An optional npts long 0/1 array to mask points in pts.

Returns: A LALInferenceKDE structure containing the KDE of the distribution.

See also: LALInferenceNewKDE()

Definition at line 109 of file LALInferenceKDE.c.

◆ LALInferenceInitKDE()

LALInferenceKDE * LALInferenceInitKDE	(	INT4	npts,
		INT4	dim
	)

Construct an empty KDE structure.

Create an empty LALInferenceKDE structure, allocated to handle the given size and dimension.

Parameters

[in]	npts	Number of samples that will be used to estimate the distribution.
[in]	dim	Number of dimensions that the probability density function will be estimated in.

Returns: An allocated, empty LALInferenceKDE structure.

See also: LALInferenceKDE, LALInferenceSetKDEBandwidth()

Definition at line 155 of file LALInferenceKDE.c.

◆ LALInferenceDestroyKDE()

void LALInferenceDestroyKDE ( LALInferenceKDE * kde )

Free an allocated KDE structure.

Frees all memory allocated for a given KDE structure.

Parameters

[in] kde The KDE structure to be freed.

See also: LALInferenceKDE, LALInferenceInitKDE

Definition at line 189 of file LALInferenceKDE.c.

◆ LALInferenceSetKDEBandwidth()

void LALInferenceSetKDEBandwidth ( LALInferenceKDE * kde )

Calculate the bandwidth and normalization factor for a KDE.

Use Scott's rule to determine the bandwidth, and corresponding normalization factor, for a KDE.

Parameters

[in] kde The kernel density estimate to estimate the bandwidth of.

Definition at line 247 of file LALInferenceKDE.c.

◆ LALInferenceKDEEvaluatePoint()

REAL8 LALInferenceKDEEvaluatePoint	(	LALInferenceKDE *	kde,
		REAL8 *	point
	)

Evaluate the (log) PDF from a KDE at a single point.

Calculate the (log) value of the probability density function estimate from a kernel density estimate at a single point.

Parameters

[in]	kde	The kernel density estimate to evaluate.
[in]	point	An array containing the point to evaluate the PDF at.

Returns: The value of the estimated probability density function at point.

Definition at line 300 of file LALInferenceKDE.c.

◆ LALInferenceDrawKDESample()

REAL8 * LALInferenceDrawKDESample	(	LALInferenceKDE *	kde,
		gsl_rng *	rng
	)

Draw a sample from a kernel density estimate.

Draw a sample from a distribution, as estimated by a kernel density estimator.

Parameters

[in]	kde	The kernel density estimate to draw point from.
[in]	rng	GSL random number generator to use.

Returns: The sample drawn from the distribution.

Definition at line 429 of file LALInferenceKDE.c.

◆ LALInferenceCholeskyDecompose()

INT4 LALInferenceCholeskyDecompose ( gsl_matrix * mat )

Compute the Cholesky decomposition of a matrix.

A wrapper for gsl_linalg_cholesky_decomp() that avoids halting if the matrix is found not to be positive-definite. This often happens when decomposing a covariance matrix that is poorly estimated due to low sample size.

Parameters

mat	The matrix to decompose (in place).

Returns: Status of call to gsl_linalg_cholesky_decomp().

Definition at line 217 of file LALInferenceKDE.c.

◆ LALInferenceComputeCovariance()

void LALInferenceComputeCovariance	(	gsl_matrix *	cov,
		gsl_matrix *	data
	)

Calculate the covariance matrix of a data set.

Calculate the covariance matrix of a data set contained in the rows of a GSL matrix.

Parameters

[out]	cov	GSL matrix containing the covariance matrix of data.
[in]	data	GSL matrix data set to compute the covariance matrix of.

Definition at line 554 of file LALInferenceKDE.c.

◆ LALInferenceComputeMean()

void LALInferenceComputeMean	(	gsl_vector *	mean,
		gsl_matrix *	data
	)

Calculate the mean of a data set.

Calculate the mean vector of a data set contained in the rows of a GSL matrix.

Parameters

[out]	mean	GSL vector containing the mean.
[in]	data	GSL matrix data set to compute the mean of.

Definition at line 528 of file LALInferenceKDE.c.

◆ LALInferenceMatrixDet()

REAL8 LALInferenceMatrixDet ( gsl_matrix * mat )

Calculate the determinant of a matrix.

Calculates the determinant of a GSL matrix.

Parameters

[in] mat The matrix to calculate the determinant of.

Returns: The determinant of mat.

Definition at line 500 of file LALInferenceKDE.c.

◆ log_add_exps()

REAL8 log_add_exps	(	REAL8 *	vals,
		INT4	size
	)

Determine the log of the sum of an array of exponentials.

Utility for calculating the log of the sum of an array of exponentials. Useful for avoiding overflows.

Parameters

[in]	vals	Array of values to be exponentiated, summed, and logged.
[in]	size	Number of elements in vals.

Returns: The log of the sum of elements in vals.

Definition at line 601 of file LALInferenceKDE.c.

Prototypes

Data Structures

Function Documentation

◆ LALInferenceNewKDE()

◆ LALInferenceNewKDEfromMat()

◆ LALInferenceInitKDE()

◆ LALInferenceDestroyKDE()

◆ LALInferenceSetKDEBandwidth()

◆ LALInferenceKDEEvaluatePoint()

◆ LALInferenceDrawKDESample()

◆ LALInferenceCholeskyDecompose()

◆ LALInferenceComputeCovariance()

◆ LALInferenceComputeMean()

◆ LALInferenceMatrixDet()

◆ log_add_exps()