#include <AdaptiveRSDiag.h>

Inheritance diagram for AdaptiveRSDiag:

Public Member Functions
	AdaptiveRSDiag (shared_ptr< MatrixXd > data, shared_ptr< Kernel > k, double lb, double eps)

	AdaptiveRSDiag (shared_ptr< MatrixXd > data, shared_ptr< Kernel > k, int samples, double lb, double eps)

int	findActualLevel (VectorXd &q, double est, double eps)

void	findRings (int strategy, double eps, VectorXd &q, int level)

void	getConstants ()
	Precompute.

void	clearSamples ()

double	vbRS ()

double	vbHBE ()

Public Member Functions inherited from AdaptiveEstimator
std::vector< double >	query (VectorXd q)

void	setMedians (int l)

Public Attributes
vector< int >	samples

vector< double >	contrib

double	lambda

double	l

Public Attributes inherited from AdaptiveEstimator
double	totalTime = 0

Protected Member Functions
std::vector< double >	evaluateQuery (VectorXd q, int level)

Additional Inherited Members
Protected Attributes inherited from AdaptiveEstimator
int	numPoints

double	gamma = 0.5

int	I

int	L = 3

std::vector< double >	mui

std::vector< int >	Mi

std::vector< double >	ti

std::vector< int >	ki

std::vector< double >	wi

double	r

std::string	EXP_STR = "exp"

Detailed Description

Diagnostic procedure implemented via adaptive random sampling.

Member Function Documentation

◆ evaluateQuery()

std::vector< double > AdaptiveRSDiag::evaluateQuery	(	VectorXd	q,
		int	level
	)

protectedvirtual

For subclasses to implement: evaluate density of query q at the given level.

Implements AdaptiveEstimator.

◆ findActualLevel()

int AdaptiveRSDiag::findActualLevel	(	VectorXd &	q,
		double	est,
		double	eps
	)

Parameters

q	query
est	final estimate from adapative sampling
eps	relative error

Returns: smallest level that we can get an esimate within (1+/-eps) of est

◆ findRings()

void AdaptiveRSDiag::findRings	(	int	strategy,
		double	eps,
		VectorXd &	q,
		int	level
	)

Set lambda and l according to different strategies

Parameters

strategy	0: S2, S3 = {}; 1: Eq(9), (10)
eps
q
level

◆ vbHBE()

double AdaptiveRSDiag::vbHBE ( )

Returns: variance upper bound for HBE

◆ vbRS()

double AdaptiveRSDiag::vbRS ( )

Returns: variance upper bound for RS

Member Data Documentation

◆ contrib

vector<double> AdaptiveRSDiag::contrib

k(q, x) for each x in S0

◆ l

double AdaptiveRSDiag::l

Eq (10) in the main paper. Used for visualization.

◆ lambda

double AdaptiveRSDiag::lambda

Eq (9) in the main paper. Used for visualization.

◆ samples

vector<int> AdaptiveRSDiag::samples

S0 in Algorithm 1, line 4

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

AdaptiveRSDiag.h
AdaptiveRSDiag.cpp

Public Member Functions

Public Attributes

Protected Member Functions

Additional Inherited Members