jsat.math

Class ExponentialMovingStatistics

java.lang.Object

jsat.math.ExponentialMovingStatistics

All Implemented Interfaces:

Serializable, Cloneable

public class ExponentialMovingStatistics
extends Object
implements Serializable, Cloneable

This class keeps track of a set of Exponential Moving statistics (the mean and standard deviation). When considering just the mean, this is often referred to as Exponential Moving Average (EMA). Similar to OnLineStatistics, this method will use fixed memory to keep an estimate of the mean and standard deviation of a stream of values. However this class will adjust to the mean and standard deviation of only recent additions, and will "forget" the contribution of earlier values. The rate of forgetting is controlled with the smoothing parameter.

Author:

Edward Raff

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
ExponentialMovingStatistics() Creates a new object for keeping an exponential estimate of the mean and variance.
ExponentialMovingStatistics(double smoothing) Creates a new object for keeping an exponential estimate of the mean and variance
ExponentialMovingStatistics(double smoothing, double mean, double variance) Creates a new object for keeping an exponential estimate of the mean and variance

Method Summary

Modifier and Type	Method and Description
void	add(double x) Adds the given data point to the statistics
double	getMean()
double	getSmoothing()
double	getStandardDeviation()
double	getVariance()
void	setSmoothing(double smoothing) Sets the smoothing parameter value to use.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

ExponentialMovingStatistics

public ExponentialMovingStatistics()

Creates a new object for keeping an exponential estimate of the mean and variance. Uses a relatively low smoothing factor of 0.1

ExponentialMovingStatistics

public ExponentialMovingStatistics(double smoothing)

Creates a new object for keeping an exponential estimate of the mean and variance

Parameters:

smoothing - the smoothing parameter to use

ExponentialMovingStatistics

public ExponentialMovingStatistics(double smoothing,
                                   double mean,
                                   double variance)

Creates a new object for keeping an exponential estimate of the mean and variance

Parameters:

smoothing - the smoothing parameter to use

mean - an initial mean. May be NaN to indicate no initial mean.

variance - an initial variance. May be NaN to indicate no initial mean.

Method Detail

setSmoothing

public void setSmoothing(double smoothing)

Sets the smoothing parameter value to use. Must be in the range (0, 1]. Changing this value will impact how quickly the statistics adapt to changes, with larger values increasing rate of change and smaller values decreasing it.

Parameters:

smoothing - the smoothing value to use

getSmoothing

public double getSmoothing()

Returns:

the smoothing parameter in use

add

public void add(double x)

Adds the given data point to the statistics

Parameters:

x - the new value to add to the moving statistics

getMean

public double getMean()

Returns:

estimate of the moving mean

getVariance

public double getVariance()

Returns:

the estimate of moving variance

getStandardDeviation

public double getStandardDeviation()

Returns:

the estimate of moving standard deviation