USGS

Isis 3.0 Object Programmers' Reference

Home

Isis::Statistics Class Reference
[Statistics]

#include <Statistics.h>

Inheritance diagram for Isis::Statistics:

Inheritance graph
[legend]
Collaboration diagram for Isis::Statistics:

Collaboration graph
[legend]
List of all members.

Detailed Description

This class is used to accumulate statistics on double arrays.

This class is used to accumulate statistics on double arrays. In particular, it is highly useful for obtaining statistics on cube data. Parameters which can be computed are 1) average, 2) standard deviation, 3) variance, 4) minimum, 5) maximum and 6) various counts of valid and/or special pixels.

The following example shows a simple set up and usage of the Statistics class to calculate the average of a set of values:

   Statistics myStats ;
   double myData [] = { 1.0, 3.0, 2.4, 7.5 } ;

   myStats.AddData (myData, 4) ;
   double myAverage = myStats.Average () ;
   cout << "The average of the data is " << myAverage << endl ;

For an example of how the Statistics object is used in Isis, see the Histogram object (inherits from Statistics) and the stats application, stats.cpp (uses the Statistics child class Histogram).

Author:
Jeff Anderson - 2002-05-06

For internal use only.

History:
2002-05-08 Jeff Anderson - Added Chebyshev and Best minimum/maximum methods.
History:
2004-05-11 Jeff Anderson - Moved Reset, AddData and RemoveData methods into public space.
History:
2004-06-28 Jeff Anderson - Added Sum and SumSquare methods.
History:
2005-02-17 Deborah Lee Soltesz - Modified file to support Doxygen documentation.
History:
2005-05-23 Jeff Anderson - Changed to support 2GB+ files
History:
2006-02-15 Jacob Danton - Added Valid Range options/methods
History:
2006-03-10 Jacob Danton - Added Z-score method
History:
2007-01-18 Robert Sucharski - Added AddData method for a single double value
History:
2008-05-06 Steven Lambright - Added AboveRange, BelowRange methods
Todo:
2005-02-07 Deborah Lee Soltesz - add example using cube data to the class documentation

Definition at line 76 of file Statistics.h.

Public Member Functions

 Statistics ()
 Constructs an IsisStats object with accumulators and counters set to zero.
 ~Statistics ()
 Destroys the IsisStats object.
void Reset ()
 Reset all accumulators and counters to zero.
void AddData (const double *data, const unsigned int count)
 Add an array of doubles to the accumulators and counters.
void AddData (const double data)
 Add a double to the accumulators and counters.
void RemoveData (const double *data, const unsigned int count)
 Remove an array of doubles from the accumulators and counters.
void RemoveData (const double data)
void SetValidRange (const double minimum=Isis::ValidMinimum, const double maximum=Isis::ValidMaximum)
double ValidMinimum () const
double ValidMaximum () const
bool InRange (const double value)
bool AboveRange (const double value)
bool BelowRange (const double value)
double Average () const
 Computes and returns the average.
double StandardDeviation () const
 Computes and returns the standard deviation.
double Variance () const
 Computes and returns the variance.
double Minimum () const
 Returns the absolute minimum double found in all data passed through the AddData method.
double Maximum () const
 Returns the absolute maximum double found in all data passed through the AddData method.
double ChebyshevMinimum (const double percent=99.5) const
 This method returns a minimum such that X percent of the data will fall with K standard deviations of the average (Chebyshev's Theorem).
double ChebyshevMaximum (const double percent=99.5) const
 This method returns a maximum such that X percent of the data will fall with K standard deviations of the average (Chebyshev's Theorem).
double BestMinimum (const double percent=99.5) const
 This method returns the better of the absolute minimum or the Chebyshev minimum.
double BestMaximum (const double percent=99.5) const
 This method returns the better of the absolute maximum or the Chebyshev maximum.
double ZScore (const double value) const
 This method returns the better of the z-score of the given value.
BigInt TotalPixels () const
 Returns the total number of pixels processed (valid and invalid).
BigInt ValidPixels () const
 Returns the total number of valid pixels processed.
BigInt OverRangePixels () const
 Returns the total number of pixels over the valid range encountered.
BigInt UnderRangePixels () const
 Returns the total number of pixels under the valid range encountered.
BigInt NullPixels () const
 Returns the total number of NULL pixels encountered.
BigInt LisPixels () const
 Returns the total number of low instrument saturation (LIS) pixels encountered.
BigInt LrsPixels () const
 Returns the total number of low representation saturation (LRS) pixels encountered.
BigInt HisPixels () const
 Returns the total number of high instrument saturation (HIS) pixels encountered.
BigInt HrsPixels () const
 Returns the total number of high representation saturation (HRS) pixels encountered.
BigInt OutOfRangePixels () const
 Returns the total number of pixels outside of the valid range encountered.
double Sum () const
 Returns the sum of all the data.
double SumSquare () const
 Returns the sum of all the squared data.

Private Attributes

double p_sum
 Sum accumulator.
double p_sumsum
 Sum-squared accumulator.
double p_minimum
 Minimum double value encountered.
double p_maximum
 Maximum double value encountered.
double p_validMinimum
 Minimum valid pixel value.
double p_validMaximum
 Maximum valid pixel value.
BigInt p_totalPixels
 Count of total pixels processed.
BigInt p_validPixels
 Count of valid pixels (non-special) processed.
BigInt p_nullPixels
 Count of null pixels processed.
BigInt p_lrsPixels
 Count of low instrument saturation pixels processed.
BigInt p_lisPixels
 Count of low representation saturation pixels processed.
BigInt p_hrsPixels
 Count of high instrument saturation pixels processed.
BigInt p_hisPixels
 Count of high instrument representation pixels processed.
BigInt p_underRangePixels
 Count of pixels less than the valid range.
BigInt p_overRangePixels
 Count of pixels greater than the valid range.
bool p_removedData
 Indicates the RemoveData method was called which implies p_minimum and p_maximum are invalid.


Constructor & Destructor Documentation

Isis::Statistics::Statistics (  ) 

Constructs an IsisStats object with accumulators and counters set to zero.

Definition at line 30 of file Statistics.cpp.

References Reset(), and SetValidRange().

Referenced by Isis::Histogram::InitializeFromCube().

Isis::Statistics::~Statistics (  ) 

Destroys the IsisStats object.

Definition at line 54 of file Statistics.cpp.


Member Function Documentation

void Isis::Statistics::AddData ( const double  data  )  [inline]

Add a double to the accumulators and counters.

This method can be invoked multiple times (for example: once for each pixel in a cube) before obtaining statistics.

Parameters:
data The data to be added to the data set used for statistical calculations.

Definition at line 92 of file Statistics.h.

References AboveRange(), InRange(), Isis::IsHisPixel(), Isis::IsHrsPixel(), Isis::IsLisPixel(), Isis::IsLrsPixel(), Isis::IsNullPixel(), Isis::IsValidPixel(), p_hisPixels, p_hrsPixels, p_lisPixels, p_lrsPixels, p_maximum, p_minimum, p_nullPixels, p_overRangePixels, p_sum, p_sumsum, p_totalPixels, p_underRangePixels, and p_validPixels.

void Isis::Statistics::AddData ( const double *  data,
const unsigned int  count 
)

Add an array of doubles to the accumulators and counters.

This method can be invoked multiple times (for example: once for each line in a cube) before obtaining statistics.

Parameters:
data The data to be added to the data set used for statistical calculations.
count The number of elements in the incoming data to be added.

Reimplemented in Isis::Histogram.

Definition at line 66 of file Statistics.cpp.

References value.

Referenced by Isis::MultivariateStatistics::AddData(), Isis::Histogram::AddData(), Isis::BundleAdjust::AddPartials(), Qisis::PlotTool::changePlot(), Isis::AutoReg::ComputeChipZScore(), Isis::Histogram::InitializeFromCube(), Isis::StandardDeviationOperator::Interest(), Isis::AutoReg::Reduce(), Isis::ReseauDistortionMap::SetFocalPlane(), Qisis::VisualDisplay::setPixelData(), Isis::ReseauDistortionMap::SetUndistortedFocalPlane(), Isis::Cube::Statistics(), Isis::Chip::Statistics(), and Isis::Gruen::TestConstraints().

double Isis::Statistics::Average (  )  const

Computes and returns the average.

If there are no valid pixels, then NULL8 is returned.

Returns:
The Average

Definition at line 140 of file Statistics.cpp.

References Isis::NULL8, p_sum, and p_validPixels.

Referenced by Isis::OverlapNormalization::AddOverlap(), Qisis::HistogramTool::changePlot(), ChebyshevMaximum(), ChebyshevMinimum(), Isis::MultivariateStatistics::Covariance(), Isis::operator<<(), Qisis::VisualDisplay::paintPixmap(), Isis::AutoReg::Reduce(), Isis::Histogram::Skew(), Isis::Gruen::StatsLog(), and ZScore().

double Isis::Statistics::BestMaximum ( const double  percent = 99.5  )  const

This method returns the better of the absolute maximum or the Chebyshev maximum.

The better value is considered the value closest to the mean.

Parameters:
percent The probability that the maximum is within K standard deviations of the mean (Used to compute the Chebyshev maximum). Default value = 99.5.
Returns:
Best of absolute and Chebyshev maximums
See also:
Statistics::Maximum Statistics::ChebyshevMaximum

Definition at line 401 of file Statistics.cpp.

References ChebyshevMaximum(), Maximum(), Isis::NULL8, and p_validPixels.

Referenced by Qisis::MosaicItem::getStretch(), Isis::Histogram::InitializeFromCube(), Qisis::VisualDisplay::setPixelData(), Qisis::StretchTool::stretchBuffer(), and Qisis::StretchTool::stretchEntireCube().

double Isis::Statistics::BestMinimum ( const double  percent = 99.5  )  const

This method returns the better of the absolute minimum or the Chebyshev minimum.

The better value is considered the value closest to the mean.

Parameters:
percent The probability that the minimum is within K standard deviations of the mean (Used to compute the Chebyshev minimum). Default value = 99.5.
Returns:
Best of absolute and Chebyshev minimums
See also:
Statistics::Minimum Statistics::ChebyshevMinimum

Definition at line 379 of file Statistics.cpp.

References ChebyshevMinimum(), Minimum(), Isis::NULL8, and p_validPixels.

Referenced by Qisis::MosaicItem::getStretch(), Isis::Histogram::InitializeFromCube(), Qisis::VisualDisplay::setPixelData(), Qisis::StretchTool::stretchBuffer(), and Qisis::StretchTool::stretchEntireCube().

double Isis::Statistics::ChebyshevMaximum ( const double  percent = 99.5  )  const

This method returns a maximum such that X percent of the data will fall with K standard deviations of the average (Chebyshev's Theorem).

It can be used to obtain a minimum that does not include statistical outliers.

Parameters:
percent The probability that the maximum is within K standard deviations of the mean. Default value = 99.5.
Returns:
maximum value excluding statistical outliers
Exceptions:
Isis::iException::Message 

Definition at line 354 of file Statistics.cpp.

References _FILEINFO_, Average(), Isis::iException::Message(), Isis::NULL8, p_validPixels, and StandardDeviation().

Referenced by BestMaximum().

double Isis::Statistics::ChebyshevMinimum ( const double  percent = 99.5  )  const

This method returns a minimum such that X percent of the data will fall with K standard deviations of the average (Chebyshev's Theorem).

It can be used to obtain a minimum that does not include statistical outliers.

Parameters:
percent The probability that the minimum is within K standard deviations of the mean. Default value = 99.5.
Returns:
Minimum value (excluding statistical outliers)
Exceptions:
Isis::iException::Message 

Definition at line 328 of file Statistics.cpp.

References _FILEINFO_, Average(), Isis::iException::Message(), Isis::NULL8, p_validPixels, and StandardDeviation().

Referenced by BestMinimum().

BigInt Isis::Statistics::HisPixels (  )  const

Returns the total number of high instrument saturation (HIS) pixels encountered.

Returns:
The number of HIS pixels (data) processed

Definition at line 289 of file Statistics.cpp.

References p_hisPixels.

BigInt Isis::Statistics::HrsPixels (  )  const

Returns the total number of high representation saturation (HRS) pixels encountered.

Returns:
The number of HRS pixels (data) processed

Definition at line 299 of file Statistics.cpp.

References p_hrsPixels.

BigInt Isis::Statistics::LisPixels (  )  const

Returns the total number of low instrument saturation (LIS) pixels encountered.

Returns:
The number of LIS pixels (data) processed

Definition at line 269 of file Statistics.cpp.

References p_lisPixels.

BigInt Isis::Statistics::LrsPixels (  )  const

Returns the total number of low representation saturation (LRS) pixels encountered.

Returns:
The number of LRS pixels (data) processed

Definition at line 279 of file Statistics.cpp.

References p_lrsPixels.

double Isis::Statistics::Maximum (  )  const

Returns the absolute maximum double found in all data passed through the AddData method.

If there are no valid pixels, then NULL8 is returned.

Returns:
Current maximum value in data set
Exceptions:
Isis::iException::Message The data set is blank, so the maximum is invalid.

Definition at line 202 of file Statistics.cpp.

References _FILEINFO_, Isis::iException::Message(), Isis::NULL8, p_maximum, p_removedData, and p_validPixels.

Referenced by BestMaximum(), Qisis::PlotTool::changePlot(), Qisis::HistogramTool::changePlot(), Isis::AutoReg::ComputeChipZScore(), Isis::GaussianStretch::GaussianStretch(), Qisis::VisualDisplay::paintPixmap(), Qisis::HistogramWidget::setHistogram(), Qisis::StretchTool::setStretch(), Isis::Gruen::StatsLog(), Qisis::StretchTool::stretchBuffer(), Qisis::StretchTool::stretchEntireCube(), Qisis::StretchTool::updateTool(), and ZScore().

double Isis::Statistics::Minimum (  )  const

Returns the absolute minimum double found in all data passed through the AddData method.

If there are no valid pixels, then NULL8 is returned.

Returns:
Current minimum value in data set.
Exceptions:
Isis::iException::Message The data set is blank, so the minimum is invalid.

Definition at line 182 of file Statistics.cpp.

References _FILEINFO_, Isis::iException::Message(), Isis::NULL8, p_minimum, p_removedData, and p_validPixels.

Referenced by BestMinimum(), Qisis::PlotTool::changePlot(), Qisis::HistogramTool::changePlot(), Isis::AutoReg::ComputeChipZScore(), Isis::GaussianStretch::GaussianStretch(), Qisis::VisualDisplay::paintPixmap(), Qisis::HistogramWidget::setHistogram(), Qisis::StretchTool::setStretch(), Isis::Gruen::StatsLog(), Qisis::StretchTool::stretchBuffer(), Qisis::StretchTool::stretchEntireCube(), and Qisis::StretchTool::updateTool().

BigInt Isis::Statistics::NullPixels (  )  const

Returns the total number of NULL pixels encountered.

Returns:
The number of NULL pixels (data) processed

Definition at line 259 of file Statistics.cpp.

References p_nullPixels.

BigInt Isis::Statistics::OutOfRangePixels (  )  const

Returns the total number of pixels outside of the valid range encountered.

Returns:
The number of Out of Range pixels (data) processed

Definition at line 309 of file Statistics.cpp.

References p_overRangePixels, and p_underRangePixels.

BigInt Isis::Statistics::OverRangePixels (  )  const

Returns the total number of pixels over the valid range encountered.

Returns:
The number of pixels less than the ValidMaximum() processed

Definition at line 240 of file Statistics.cpp.

References p_overRangePixels.

void Isis::Statistics::RemoveData ( const double *  data,
const unsigned int  count 
)

Remove an array of doubles from the accumulators and counters.

Note that is invalidates the absolute minimum and maximum. They will no longer be usable.

Parameters:
data The data to be removed from data set used for statistical calculations.
count The number of elements in the data to be removed.
Exceptions:
Isis::iException::Message RemoveData is trying to remove data that doesn't exist.

Reimplemented in Isis::Histogram.

Definition at line 88 of file Statistics.cpp.

References _FILEINFO_, AboveRange(), InRange(), Isis::IsHisPixel(), Isis::IsHrsPixel(), Isis::IsLisPixel(), Isis::IsLrsPixel(), Isis::IsNullPixel(), Isis::IsValidPixel(), Isis::iException::Message(), p_hisPixels, p_hrsPixels, p_lisPixels, p_lrsPixels, p_nullPixels, p_overRangePixels, p_removedData, p_sum, p_sumsum, p_totalPixels, p_underRangePixels, and p_validPixels.

Referenced by RemoveData(), Isis::MultivariateStatistics::RemoveData(), and Isis::Histogram::RemoveData().

void Isis::Statistics::Reset (  ) 

Reset all accumulators and counters to zero.

Reimplemented in Isis::Histogram.

Definition at line 36 of file Statistics.cpp.

References p_hisPixels, p_hrsPixels, p_lisPixels, p_lrsPixels, p_maximum, p_minimum, p_nullPixels, p_overRangePixels, p_removedData, p_sum, p_sumsum, p_totalPixels, p_underRangePixels, and p_validPixels.

Referenced by Isis::AutoReg::Reduce(), Isis::MultivariateStatistics::Reset(), Isis::Histogram::Reset(), Isis::Gruen::resetStats(), Qisis::VisualDisplay::setLines(), Qisis::VisualDisplay::setPixelData(), Qisis::VisualDisplay::setSamples(), Isis::BundleAdjust::Solve(), and Statistics().

double Isis::Statistics::StandardDeviation (  )  const

Computes and returns the standard deviation.

If there are no valid pixels, then NULL8 is returned.

Returns:
The standard deviation

Reimplemented in Isis::GaussianDistribution.

Definition at line 151 of file Statistics.cpp.

References Isis::NULL8, p_validPixels, and Variance().

Referenced by Qisis::HistogramTool::changePlot(), ChebyshevMaximum(), ChebyshevMinimum(), Isis::MultivariateStatistics::Correlation(), Isis::StandardDeviationOperator::Interest(), Isis::operator<<(), Qisis::VisualDisplay::paintPixmap(), Isis::ReseauDistortionMap::SetFocalPlane(), Isis::ReseauDistortionMap::SetUndistortedFocalPlane(), Isis::Histogram::Skew(), Isis::Gruen::StatsLog(), and ZScore().

double Isis::Statistics::Sum (  )  const [inline]

Returns the sum of all the data.

Returns:
The sum of the data

Definition at line 156 of file Statistics.h.

References p_sum.

Referenced by Isis::MultivariateStatistics::Covariance(), and Isis::MultivariateStatistics::LinearRegression().

double Isis::Statistics::SumSquare (  )  const [inline]

Returns the sum of all the squared data.

Returns:
The sum of the squared data

Definition at line 163 of file Statistics.h.

References p_sumsum.

Referenced by Isis::MultivariateStatistics::LinearRegression(), and Isis::BundleAdjust::Solve().

BigInt Isis::Statistics::TotalPixels (  )  const

Returns the total number of pixels processed (valid and invalid).

Returns:
The number of pixels (data) processed

Definition at line 218 of file Statistics.cpp.

References p_totalPixels.

Referenced by Isis::operator<<(), Qisis::VisualDisplay::paintPixmap(), and Isis::BundleAdjust::Solve().

BigInt Isis::Statistics::UnderRangePixels (  )  const

Returns the total number of pixels under the valid range encountered.

Returns:
The number of pixels less than the ValidMinimum() processed

Definition at line 250 of file Statistics.cpp.

References p_underRangePixels.

BigInt Isis::Statistics::ValidPixels (  )  const

Returns the total number of valid pixels processed.

Only valid pixels are utilized when computing the average, standard deviation, variance, minimum and maximum.

Returns:
The number of valid pixels (data) processed

Definition at line 230 of file Statistics.cpp.

References p_validPixels.

Referenced by Isis::OverlapNormalization::AddOverlap(), Qisis::HistogramTool::changePlot(), Qisis::StretchTool::getHistogramPercent(), Isis::Histogram::InitializeFromCube(), Qisis::QHistogram::Load(), Isis::operator<<(), Isis::Histogram::Percent(), and Isis::Histogram::Skew().

double Isis::Statistics::Variance (  )  const

Computes and returns the variance.

If there are no valid pixels, then NULL8 is returned.

Returns:
The variance

For internal use only.

History:
2003-08-27 Jeff Anderson - Modified Variance method to compute using n*(n-1) instead of n*n.

Definition at line 166 of file Statistics.cpp.

References Isis::NULL8, p_sum, p_sumsum, and p_validPixels.

Referenced by Qisis::HistogramTool::changePlot(), Isis::operator<<(), and StandardDeviation().

double Isis::Statistics::ZScore ( const double  value  )  const

This method returns the better of the z-score of the given value.

The z-score is the number of standard deviations the value lies above or below the average.

Parameters:
value The value to calculate the z-score of.
Returns:
z-score

Definition at line 420 of file Statistics.cpp.

References _FILEINFO_, Average(), Maximum(), Isis::iException::Message(), and StandardDeviation().

Referenced by Isis::AutoReg::ComputeChipZScore().


Member Data Documentation

BigInt Isis::Statistics::p_hisPixels [private]

Count of high instrument representation pixels processed.

Definition at line 178 of file Statistics.h.

Referenced by AddData(), HisPixels(), RemoveData(), and Reset().

BigInt Isis::Statistics::p_hrsPixels [private]

Count of high instrument saturation pixels processed.

Definition at line 177 of file Statistics.h.

Referenced by AddData(), HrsPixels(), RemoveData(), and Reset().

BigInt Isis::Statistics::p_lisPixels [private]

Count of low representation saturation pixels processed.

Definition at line 176 of file Statistics.h.

Referenced by AddData(), LisPixels(), RemoveData(), and Reset().

BigInt Isis::Statistics::p_lrsPixels [private]

Count of low instrument saturation pixels processed.

Definition at line 175 of file Statistics.h.

Referenced by AddData(), LrsPixels(), RemoveData(), and Reset().

double Isis::Statistics::p_maximum [private]

Maximum double value encountered.

Definition at line 169 of file Statistics.h.

Referenced by AddData(), Maximum(), and Reset().

double Isis::Statistics::p_minimum [private]

Minimum double value encountered.

Definition at line 168 of file Statistics.h.

Referenced by AddData(), Minimum(), and Reset().

BigInt Isis::Statistics::p_nullPixels [private]

Count of null pixels processed.

Definition at line 174 of file Statistics.h.

Referenced by AddData(), NullPixels(), RemoveData(), and Reset().

BigInt Isis::Statistics::p_overRangePixels [private]

Count of pixels greater than the valid range.

Definition at line 180 of file Statistics.h.

Referenced by AddData(), OutOfRangePixels(), OverRangePixels(), RemoveData(), and Reset().

bool Isis::Statistics::p_removedData [private]

Indicates the RemoveData method was called which implies p_minimum and p_maximum are invalid.

Definition at line 181 of file Statistics.h.

Referenced by Maximum(), Minimum(), RemoveData(), and Reset().

double Isis::Statistics::p_sum [private]

Sum accumulator.

Definition at line 163 of file Statistics.h.

Referenced by AddData(), Average(), RemoveData(), Reset(), Sum(), and Variance().

double Isis::Statistics::p_sumsum [private]

Sum-squared accumulator.

Definition at line 167 of file Statistics.h.

Referenced by AddData(), RemoveData(), Reset(), SumSquare(), and Variance().

BigInt Isis::Statistics::p_totalPixels [private]

Count of total pixels processed.

Definition at line 172 of file Statistics.h.

Referenced by AddData(), RemoveData(), Reset(), and TotalPixels().

BigInt Isis::Statistics::p_underRangePixels [private]

Count of pixels less than the valid range.

Definition at line 179 of file Statistics.h.

Referenced by AddData(), OutOfRangePixels(), RemoveData(), Reset(), and UnderRangePixels().

double Isis::Statistics::p_validMaximum [private]

Maximum valid pixel value.

Definition at line 171 of file Statistics.h.

Referenced by AboveRange(), InRange(), SetValidRange(), and ValidMaximum().

double Isis::Statistics::p_validMinimum [private]

Minimum valid pixel value.

Definition at line 170 of file Statistics.h.

Referenced by BelowRange(), InRange(), SetValidRange(), and ValidMinimum().

BigInt Isis::Statistics::p_validPixels [private]

Count of valid pixels (non-special) processed.

Definition at line 173 of file Statistics.h.

Referenced by AddData(), Average(), BestMaximum(), BestMinimum(), ChebyshevMaximum(), ChebyshevMinimum(), Maximum(), Minimum(), RemoveData(), Reset(), StandardDeviation(), ValidPixels(), and Variance().


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