com.numericalmethod.suanshu.stats.timeseries.linear.univariate.stationaryprocess
Class MADecomposition

java.lang.Object
  com.numericalmethod.suanshu.stats.timeseries.linear.univariate.stationaryprocess.MADecomposition

public class MADecomposition
extends java.lang.Object

This class decomposes a time series into the trend, seasonal and stationary random components using the Moving Average Estimation method with symmetric window. That is,

X_t = m_t + s_t + Y_t

We have

• EY_t = 0
• s_t+d = s_t
• Σ(s_j) = 0

The R equivalent function is decompose.

See Also:

"P. J. Brockwell and R. A. Davis, "p. 23, Chapter 1.4, Elimination of both Trend and Seasonality," Time Series: Theory and Methods, 2nd ed, Springer, 2006."

Constructor Summary

MADecomposition(double[] xt, double[] MAFilter, int period)
Decompose a time series into the trend, seasonal and stationary random components using the Moving Average Estimation method.

MADecomposition(double[] xt, int period)
Decompose a periodic time series into the seasonal and stationary random components using no MA filter.

MADecomposition(double[] xt, int MAOrder, int period)
Decompose a time series into the trend, seasonal and stationary random components using the default filter.

Method Summary

double[]	getRandom() Get the estimated seasonal effect of the time series.
double[]	getSeasonal() Get the stationary random component of the time series after the trend and seasonal components are removed.
double[]	getTrend() Get the estimated trend of the time series.

Methods inherited from class java.lang.Object

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

Constructor Detail

MADecomposition

public MADecomposition(double[] xt,
                       double[] MAFilter,
                       int period)

Decompose a time series into the trend, seasonal and stationary random components using the Moving Average Estimation method.

Parameters:

xt - a time series

MAFilter - the moving average filter to smooth the time series

period - the period of the time series; if aperiodic, use 1

MADecomposition

public MADecomposition(double[] xt,
                       int MAOrder,
                       int period)

Decompose a time series into the trend, seasonal and stationary random components using the default filter.

Parameters:

xt - a time series

MAOrder - the length of the MA filter (automatically increased by 1 for even MAOrder)

period - the period of the time series; if aperiodic, use 0

See Also:

"P. J. Brockwell and R. A. Davis, "Eq. 1.4.16, Chapter 1.4, Elimination of both Trend and Seasonality," Time Series: Theory and Methods, 2nd ed, Springer, 2006."

MADecomposition

public MADecomposition(double[] xt,
                       int period)

Decompose a periodic time series into the seasonal and stationary random components using no MA filter.

Parameters:

xt - a time series

period - the period of the time series; if aperiodic, use 0

Method Detail

getTrend

public double[] getTrend()

Get the estimated trend of the time series.

Returns:

the estimated trend

getSeasonal

public double[] getSeasonal()

Get the stationary random component of the time series after the trend and seasonal components are removed.

Returns:

the stationary random component

getRandom

public double[] getRandom()

Get the estimated seasonal effect of the time series.

Returns:

the estimated seasonal effect