Accumulate mean, stddev, min, and max over a streaming data set. More...

#include <running_statistics.hpp>

Public Member Functions
Modifiers
void	reset () noexcept
	All values return to their defaults. More...

void	insert (double val) noexcept
	Add a new value to the data set. More...

Queries
size_t	samples () const noexcept
	Number of observed samples. More...

double	min () const noexcept
	Get the minimum observed value. More...

double	max () const noexcept
	Get the minimum observed value. More...

double	mean () const noexcept
	Running mean of observed samples. More...

double	total () const noexcept
	Running sum of observed samples. More...

double	variance () const noexcept
	Running (unbiased) sample variance of the observed samples. More...

double	stddev () const noexcept
	Running (unbiased) sample standard deviation of the observed samples. More...

Static Public Attributes
static constexpr auto	default_min = std::numeric_limits<double>::max()
	The default minimum value, chosen to ensure that any observed sample is less-than this. More...

static constexpr auto	default_max = std::numeric_limits<double>::lowest()
	The default maximum value, chosen to ensure that any observed sample is greater-than this. More...

Private Attributes
size_t	m_count = 0UL

double	m_min = default_min

double	m_max = default_max

double	m_mean = 0.

double	m_diff_sq = 0.

Detailed Description

Accumulate mean, stddev, min, and max over a streaming data set.

Definition at line 40 of file running_statistics.hpp.

Member Function Documentation

◆ insert()

void lbann::RunningStats::insert ( double val )

inlinenoexcept

Add a new value to the data set.

The statistics are updated immediately.

Definition at line 110 of file running_statistics.hpp.

Here is the caller graph for this function:

◆ max()

double lbann::RunningStats::max ( ) const

inlinenoexcept

Get the minimum observed value.

Note: Only valid if samples() > 0.

Definition at line 125 of file running_statistics.hpp.

Here is the caller graph for this function:

◆ mean()

double lbann::RunningStats::mean ( ) const

inlinenoexcept

Running mean of observed samples.

Definition at line 127 of file running_statistics.hpp.

Here is the caller graph for this function:

◆ min()

double lbann::RunningStats::min ( ) const

inlinenoexcept

Get the minimum observed value.

Note: Only valid if samples() > 0.

Definition at line 123 of file running_statistics.hpp.

Here is the caller graph for this function:

◆ reset()

void lbann::RunningStats::reset ( )

inlinenoexcept

All values return to their defaults.

All queries should be considered meaningless unless samples() is greater than 0.

Definition at line 108 of file running_statistics.hpp.

Here is the caller graph for this function:

◆ samples()

size_t lbann::RunningStats::samples ( ) const

inlinenoexcept

Number of observed samples.

Definition at line 121 of file running_statistics.hpp.

Here is the caller graph for this function:

◆ stddev()

double lbann::RunningStats::stddev ( ) const

inlinenoexcept

Running (unbiased) sample standard deviation of the observed samples.

Note: This is zero at least until 2 or more samples have been observed.

Definition at line 139 of file running_statistics.hpp.

Here is the call graph for this function:

Here is the caller graph for this function:

◆ total()

double lbann::RunningStats::total ( ) const

inlinenoexcept

Running sum of observed samples.

Definition at line 129 of file running_statistics.hpp.

Here is the caller graph for this function:

◆ variance()

double lbann::RunningStats::variance ( ) const

inlinenoexcept

Running (unbiased) sample variance of the observed samples.

Note: This is zero at least until 2 or more samples have been observed.

Definition at line 134 of file running_statistics.hpp.

Here is the caller graph for this function:

Member Data Documentation

◆ default_max

constexpr auto lbann::RunningStats::default_max = std::numeric_limits<double>::lowest()

static

The default maximum value, chosen to ensure that any observed sample is greater-than this.

Definition at line 50 of file running_statistics.hpp.

◆ default_min

constexpr auto lbann::RunningStats::default_min = std::numeric_limits<double>::max()

static

The default minimum value, chosen to ensure that any observed sample is less-than this.

Definition at line 46 of file running_statistics.hpp.

◆ m_count

size_t lbann::RunningStats::m_count = 0UL

private

Definition at line 101 of file running_statistics.hpp.

◆ m_diff_sq

double lbann::RunningStats::m_diff_sq = 0.

private

Definition at line 105 of file running_statistics.hpp.

◆ m_max

double lbann::RunningStats::m_max = default_max

private

Definition at line 103 of file running_statistics.hpp.

◆ m_mean

double lbann::RunningStats::m_mean = 0.

private

Definition at line 104 of file running_statistics.hpp.

◆ m_min

double lbann::RunningStats::m_min = default_min

private

Definition at line 102 of file running_statistics.hpp.

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

running_statistics.hpp

Public Member Functions

Static Public Attributes

Private Attributes

Detailed Description

Member Function Documentation

◆ insert()

◆ max()

◆ mean()

◆ min()

◆ reset()

◆ samples()

◆ stddev()

◆ total()

◆ variance()

Member Data Documentation

◆ default_max

◆ default_min

◆ m_count

◆ m_diff_sq

◆ m_max

◆ m_mean

◆ m_min