/*
* Copyright (C) 2016 Edward Raff
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see .
*/
package jsat.math;
import java.io.Serializable;
/**
* 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
* {@link 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 {@link #smoothing smoothing} parameter.
*
*
* @author Edward Raff
*/
public class ExponentialMovingStatistics implements Serializable, Cloneable
{
private double mean;
private double variance;
private double smoothing;
/**
* Creates a new object for keeping an exponential estimate of the mean and
* variance. Uses a relatively low smoothing factor of 0.1
*
*/
public ExponentialMovingStatistics()
{
this(0.1);
}
/**
* Creates a new object for keeping an exponential estimate of the mean and
* variance
*
* @param smoothing the {@link #smoothing smoothing} parameter to use
*/
public ExponentialMovingStatistics(double smoothing)
{
this(smoothing, Double.NaN, 0);
}
/**
* Creates a new object for keeping an exponential estimate of the mean and
* variance
*
* @param smoothing the {@link #smoothing smoothing} parameter to use
* @param mean an initial mean. May be {@link Double#NaN NaN} to indicate no
* initial mean.
* @param variance an initial variance. May be {@link Double#NaN NaN} to
* indicate no initial mean.
*/
public ExponentialMovingStatistics(double smoothing, double mean, double variance)
{
this.mean = mean;
this.variance = variance;
setSmoothing(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.
*
* @param smoothing the smoothing value to use
*/
public void setSmoothing(double smoothing)
{
if (smoothing <= 0 || smoothing > 1 || Double.isNaN(smoothing))
throw new IllegalArgumentException("Smoothing must be in (0, 1], not " + smoothing);
this.smoothing = smoothing;
}
/**
*
* @return the smoothing parameter in use
*/
public double getSmoothing()
{
return smoothing;
}
/**
* Adds the given data point to the statistics
*
* @param x the new value to add to the moving statistics
*/
public void add(double x)
{
if (Double.isNaN(mean))//fist case
{
mean = x;
variance = 0;
}
else//general case
{
//first update stnd deviation
variance = (1-smoothing)*(variance + smoothing*Math.pow(x-mean, 2));
mean = (1-smoothing)*mean + smoothing*x;
}
}
/**
*
* @return estimate of the moving mean
*/
public double getMean()
{
return mean;
}
/**
*
* @return the estimate of moving variance
*/
public double getVariance()
{
return variance;
}
/**
*
* @return the estimate of moving standard deviation
*/
public double getStandardDeviation()
{
return Math.sqrt(getVariance()+1e-13);
}
}
Ads help maintain this website.