com.numericalmethod.suanshu.signalprocessing.filter
Class MovingAverage

java.lang.Object
  com.numericalmethod.suanshu.signalprocessing.filter.MovingAverage

All Implemented Interfaces:

Filter

public class MovingAverage
extends java.lang.Object
implements Filter

This class applies a linear filtering to a univariate time series using Moving Average estimation.

 y[i] = f[1]*x[i+o] + … + f[p]*x[i+o-(p-1)]
 

o is the offset, depending on whether only past values or both past and future values centered around lag 0 are used.

When a symmetric window is used and the filter length is even, more of the filter is forward in time than backward.

The R equivalent function is filter.

Nested Class Summary

static class

MovingAverage.Side the types of moving average filtering available

Field Summary

MovingAverage.Side

side the data window

Constructor Summary

MovingAverage(double[] filter)
Create a MovingAverage filter using a symmetric window.

MovingAverage(double[] filter, MovingAverage.Side side)
Create a MovingAverage filter.

Method Summary

double[]

filtering(double[] Xt) Get the filtered signals.

Methods inherited from class java.lang.Object

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

Field Detail

side

public final MovingAverage.Side side

the data window

Constructor Detail

MovingAverage

public MovingAverage(double[] filter,
                     MovingAverage.Side side)

Create a MovingAverage filter.

Parameters:

filter - the filter coefficients in reverse time order

side - specify the data window to use

MovingAverage

public MovingAverage(double[] filter)

Create a MovingAverage filter using a symmetric window.

Parameters:

filter - the filter coefficients in reverse time order

Method Detail

filtering

public double[] filtering(double[] Xt)

Description copied from interface: Filter

Get the filtered signals.

Specified by:

filtering in interface Filter

Parameters:

Xt - the input signals

Returns:

the filtered signals