com.github.servicenow.ds.stats.stl

Class SeasonalTrendLoess

java.lang.Object

com.github.servicenow.ds.stats.stl.SeasonalTrendLoess

public class SeasonalTrendLoess
extends java.lang.Object

Java implementation of the Seasonal-Trend-Loess algorithm for evenly spaced data. This is basically a direct port of the RATFOR from the netlib stl package.

Created by Jim Crotinger on 18-Apr-2016.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	SeasonalTrendLoess.Builder Builder class for SeasonalTrendLoess decomposition
static class	SeasonalTrendLoess.Decomposition Simple class to hold the results of the STL decomposition.

Method Summary

Modifier and Type	Method and Description
SeasonalTrendLoess.Decomposition	decompose() Decompose the input data into the seasonal, trend and residual components.
static SeasonalTrendLoess.Decomposition	performPeriodicDecomposition(double[] data, int periodicity) Factory method to perform a non-robust STL decomposition enforcing strict periodicity.
static SeasonalTrendLoess.Decomposition	performRobustPeriodicDecomposition(double[] data, int periodicity) Factory method to perform a (somewhat) robust STL decomposition enforcing strict periodicity.
java.lang.String	toString()

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, wait, wait, wait

Method Detail

performPeriodicDecomposition

public static SeasonalTrendLoess.Decomposition performPeriodicDecomposition(double[] data,
                                                                            int periodicity)

Factory method to perform a non-robust STL decomposition enforcing strict periodicity.

Meant for diagnostic purposes only.

Parameters:

data - the data to analyze

periodicity - the (suspected) periodicity of the data

Returns:

SeasonalTrendLoess object with the decomposition already performed.

performRobustPeriodicDecomposition

public static SeasonalTrendLoess.Decomposition performRobustPeriodicDecomposition(double[] data,
                                                                                  int periodicity)

Factory method to perform a (somewhat) robust STL decomposition enforcing strict periodicity.

Meant for diagnostic purposes only.

Parameters:

data - the data to analyze

periodicity - the (suspected) periodicity of the data

Returns:

SeasonalTrendLoess object with the decomposition already performed.

decompose

public SeasonalTrendLoess.Decomposition decompose()

Decompose the input data into the seasonal, trend and residual components.

Returns:

the STL decomposition

toString

public java.lang.String toString()

Overrides:

toString in class java.lang.Object