ipsdk::math::BaseRobustEstimator Class Reference

Base class for parametric robust estimation from data set. More...

#include <BaseRobustEstimator.h>

Inheritance diagram for ipsdk::math::BaseRobustEstimator:

Public Member Functions
bool	isInit () const
	check for object initialization
virtual ipReal64	getMaxOutlierRatio () const
	retrieve maximum outlier ratio supported by underlying algorithm
ipReal64	getExpectedOutlierRatio () const
	retrieve expected outlier ratio associated to data
void	setExpectedOutlierRatio (const ipReal64 expectedOutlierRatio)
	selection of expected outlier ratio associated to data More...
void	setSuccessRatio (const ipReal64 successRatio)
	selection of algorithm statistical success ratio parameter More...
ipReal64	getSuccessRatio () const
	retrieve algorithm statistical success ratio parameter
void	setDataPartition (const DataPartitionPtr &pDataPartition)
	select data partition object type More...
const BaseDataPartition &	getDataPartition () const
	retrieve data partition object
const BaseEstimator &	getEstimator () const
	retrieve estimator associated to object More...
ipUInt32	getNbMinDataForRobustEstimation () const
	retrieve minimum number of data to proceed to robust estimation More...
ipUInt32	getNbNeededRuns () const
	retrieve number of needed estimator run for robust estimation process More...
bool	isEstimationDone () const
	check whether estimation has been proceeded More...
BoolResult	computeEstimation (const DataSet &dataSet)
	method allowing to estimate parameters from a given data set More...
const ParameterType &	getEstimParams () const
	retrieve estimated parameters More...
const OutlierFlags &	getOutlierFlags () const
	retrieve estimated outlier flags More...
ipReal64	getEstimOutlierThreshold () const
	retrieve outlier threshold estimated during processing (data associated to residuals with absolute value greater or equal to this value are considered as outliers) More...
void	clearEstimation ()
	clear of last estimation data
Constructors and destructor
	BaseRobustEstimator ()
virtual	~BaseRobustEstimator ()=0

Protected Member Functions
void	initBase (const EstimatorPtr &pEstimator)
	base class initialization method More...
virtual ipReal64	computeOutlierThreshold (const ipReal64 bestScore, const ipUInt32 nbData, const ipUInt32 nbMinRequestedData) const =0
virtual ipReal64	computeScore (const ResidualsSet &curResiduals)=0
	method used in derived classes to compute score associated to an estimation
virtual void	clearDerived ()
	clear of data into derived class

Protected Attributes
ipReal64	_expectedOutlierRatio
ipReal64	_successRatio
DataPartitionPtr	_pDataPartition
ipUInt32	_nbNeededRuns
EstimatorPtr	_pEstimator
	shared pointer to used estimator
ParameterType	_estimParams
	parameters estimated by robust algorithm
OutlierFlags	_outlierFlags
	outlier flag collection associated to input data
ipReal64	_estimOutlierThreshold
	outlier threshold computed during processing

Detailed Description

Base class for parametric robust estimation from data set.

Author

E.Noirfalise

Date

2017/05/23

Copyright

2017 Reactiv'IP, all right reserved.

Member Function Documentation

setExpectedOutlierRatio()

void ipsdk::math::BaseRobustEstimator::setExpectedOutlierRatio

(

const ipReal64

expectedOutlierRatio

)

selection of expected outlier ratio associated to data

This parameter control outlier detection process. User should not modify it without a good knowledge of underlying algorithm.

Note

Default value : 0.4 (0.4 stands for 40% of outliers)

Exceptions

ipsdk::math::IPSDKMathException

if expectedOutlierRatio < 0 || expectedOutlierRatio >= getMaxOutlierRatio()

setSuccessRatio()

void ipsdk::math::BaseRobustEstimator::setSuccessRatio

(

const ipReal64

successRatio

)

selection of algorithm statistical success ratio parameter

This parameter control outlier detection process. User should not modify it without a good knowledge of underlying algorithm.

Note

Default value : 0.99

Exceptions

ipsdk::math::IPSDKMathException

if successRatio <= 0 || successRatio >= 1

setDataPartition()

void ipsdk::math::BaseRobustEstimator::setDataPartition

(

const DataPartitionPtr &

pDataPartition

)

select data partition object type

Exceptions

ipsdk::math::IPSDKMathException

if pDataPartition.get() == 0

Note

Default value : AdaptiveGridPartition

getEstimator()

const BaseEstimator& ipsdk::math::BaseRobustEstimator::getEstimator

(

)

const

retrieve estimator associated to object

Exceptions

ipsdk::math::IPSDKMathException

if isInit() == false

getNbMinDataForRobustEstimation()

ipUInt32 ipsdk::math::BaseRobustEstimator::getNbMinDataForRobustEstimation

(

)

const

retrieve minimum number of data to proceed to robust estimation

Exceptions

ipsdk::math::IPSDKMathException

if isInit() == false

getNbNeededRuns()

ipUInt32 ipsdk::math::BaseRobustEstimator::getNbNeededRuns

(

)

const

retrieve number of needed estimator run for robust estimation process

Exceptions

ipsdk::math::IPSDKMathException

if isInit() == false

isEstimationDone()

bool ipsdk::math::BaseRobustEstimator::isEstimationDone

(

)

const

check whether estimation has been proceeded

Exceptions

ipsdk::math::IPSDKMathException

if isInit() == false

computeEstimation()

BoolResult ipsdk::math::BaseRobustEstimator::computeEstimation

(

const DataSet &

dataSet

)

method allowing to estimate parameters from a given data set

Exceptions

ipsdk::math::IPSDKMathException

if isInit() == false

getEstimParams()

const ParameterType& ipsdk::math::BaseRobustEstimator::getEstimParams

(

)

const

retrieve estimated parameters

Exceptions

ipsdk::math::IPSDKMathException

if isEstimationDone() == false

getOutlierFlags()

const OutlierFlags& ipsdk::math::BaseRobustEstimator::getOutlierFlags

(

)

const

retrieve estimated outlier flags

Exceptions

ipsdk::math::IPSDKMathException

if isEstimationDone() == false

getEstimOutlierThreshold()

ipReal64 ipsdk::math::BaseRobustEstimator::getEstimOutlierThreshold

(

)

const

retrieve outlier threshold estimated during processing (data associated to residuals with absolute value greater or equal to this value are considered as outliers)

Exceptions

ipsdk::math::IPSDKMathException

if isEstimationDone() == false

initBase()

void ipsdk::math::BaseRobustEstimator::initBase ( const EstimatorPtr &  pEstimator )

protected

base class initialization method

Exceptions

ipsdk::math::IPSDKMathException	if pEstimator.get() == 0
ipsdk::math::IPSDKMathException	if pEstimator->isInit() == false

computeOutlierThreshold()

virtual ipReal64 ipsdk::math::BaseRobustEstimator::computeOutlierThreshold ( const ipReal64  bestScore, const ipUInt32  nbData, const ipUInt32  nbMinRequestedData  ) const

protectedpure virtual

compute outlier threshold (data associated to residuals with absolute value greater or equal to this value are considered as outliers)

Implemented in ipsdk::math::MLESACRobustEstimator, ipsdk::math::LMSRobustEstimator, ipsdk::math::MSACRobustEstimator, and ipsdk::math::RANSACRobustEstimator.

Member Data Documentation

_expectedOutlierRatio

ipReal64 ipsdk::math::BaseRobustEstimator::_expectedOutlierRatio

protected

expected outlier ratio associated to data

This parameter control outlier detection process. User should not modify it without a good knowledge of underlying algorithm.

Note

Default value : 0.4 (40% of outliers)

_successRatio

ipReal64 ipsdk::math::BaseRobustEstimator::_successRatio

protected

algorithm statistical success ratio parameter

This parameter control outlier detection process. User should not modify it without a good knowledge of underlying algorithm.

Note

Default value : 0.99

_pDataPartition

DataPartitionPtr ipsdk::math::BaseRobustEstimator::_pDataPartition

protected

shared pointer to used bucket algorithm

Note

Default value : AdaptiveGridPartition

_nbNeededRuns

ipUInt32 ipsdk::math::BaseRobustEstimator::_nbNeededRuns

protected

number of theorical runs for estimator to ensure success ratio probability of getting a sample without outlier

---

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

• /media/reactivipdev/Nightly/IPSDK_Patch/IPSDK_4_1/IPSDK_Nightly/include/public/IPSDKMath/Estimation/BaseRobustEstimator.h