Go to the previous, next section.

CNConfidence -- Evaluation with Confidence Intervals

SYNOPSIS

#include <CNCL/Confidence.h>

TYPE

CN_CONFIDENCE

BASE CLASSES

CNStatistics

DERIVED CLASSES

None

RELATED CLASSES

CNMoments CNMomentsTime

DESCRIPTION

The CNConfidence class yields the moments of an input sequence:

Constructors:

CNConfidence(CNParam *param)
CNConfidence(const char* aName = NIL, const char* aDescription = NIL);
Initializes a CNConfidence evaluation. Optionally, you can specify aName and aDescription of the statistical evaluator.

In addition to the member functions required by CNCL and CNStatistics, CNConfidence provides:

virtual double mean() const;
Returns mean of the input values.

double variance() const;
Returns variance.

double M_2() const;
Returns 2nd moment.

double M_3() const;
Returns 3rd moment.

double Z_3() const;
Returns 3rd central moment.

double skewness() const;
Returns skewness.

double relative_variance() const;
Returns relative variance (squared coefficient of variation).

double relative_deviation() const;
Returns relative deviation (coefficient of variation).

double z_level(double conf) const;
Returns Z-level (argument) of error function for given confidence level (inverse error function).

double err_level(double z_level) const;
Returns value of error function to given (Z-level) argument.

double conf_level(double z_level) const;
Returns value of confidence probability to given (Z-level) argument.

double low_conf_bound(double conf) const;
Returns lower confidence boundary to given confidence level. The estimator is determined by mean().

double hi_conf_bound(double conf) const;
Returns upper confidence boundary to given confidence level. The estimator is determined by mean().

In general the confidence level conf = 0.95 is quite useful. It means that in 1 out of 20 measurements the "true" value is within the confidence interval determined by hi_conf_bound() and low_conf_bound().

Go to the previous, next section.