com.github.signaflo.timeseries.operators

Class LagPolynomial

java.lang.Object

com.github.signaflo.timeseries.operators.LagPolynomial

public class LagPolynomial
extends Object

Represents a polynomial in the lag operator. See Harvey's Forecasting, structural time series model and the Kalman filter, (1989, equation 2.1.3), or the wiki entry. The polynomial is taken in the lag operator, but is algebraically equivalent to a real or complex polynomial.

Method Summary

Modifier and Type	Method and Description
double	apply(TimeSeries timeSeries, int index) Apply this lag polynomial to a time series at the given index.
double	apply(TimeSeries timeSeries, OffsetDateTime dateTime) Apply this lag polynomial to a time series at the given date-time.
static LagPolynomial	autoRegressive(double... parameters) Create and return a new autoregressive lag polynomial.
static LagPolynomial	differences(int d) Create and return a new lag polynomial representing an arbitrary number of differences.
boolean	equals(Object obj)
static LagPolynomial	firstDifference() Create and return a new lag polynomial representing the first difference operator.
int	hashCode()
double[]	inverseParams() Return the additive inverse of the parameters of this lag polynomial.
static LagPolynomial	movingAverage(double... parameters) Create and return a new moving average lag polynomial.
double[]	parameters() Return the parameters of this lag polynomial.
static LagPolynomial	seasonalDifferences(int seasonalLag, int D) Create and return a new lag polynomial representing an arbitrary number of seasonal differences.
double	solve(double[] timeSeries, int index) Apply this lag polynomial to the series at the given index, then solve for the value of the series at that index.
double	solve(TimeSeries timeSeries, int index) Apply this lag polynomial to the series at the given index, then solve for the value of the series at that index.
double	solve(TimeSeries timeSeries, OffsetDateTime dateTime) Apply this lag polynomial to the series at the given date-time, then solve for the value of the series at that date-time.
LagPolynomial	times(LagPolynomial other) Multiply this polynomial by another lag polynomial and return the result in a new lag polynomial.
String	toString()

Methods inherited from class java.lang.Object

getClass, notify, notifyAll, wait, wait, wait

Method Detail

firstDifference

public static LagPolynomial firstDifference()

Create and return a new lag polynomial representing the first difference operator.

Returns:

a new lag polynomial representing the first difference operator.

seasonalDifferences

public static LagPolynomial seasonalDifferences(int seasonalLag,
                                                int D)

Create and return a new lag polynomial representing an arbitrary number of seasonal differences.

Parameters:

seasonalLag - the period of seasonality; the lag to use for the operator.

D - the number of seasonal differences.

Returns:

a new lag polynomial representing an arbitrary number of seasonal differences.

differences

public static LagPolynomial differences(int d)

Create and return a new lag polynomial representing an arbitrary number of differences.

Parameters:

d - the number of differences. An integer greater than or equal to 0.

Returns:

a new lag polynomial representing an arbitrary number of differences.

movingAverage

public static LagPolynomial movingAverage(double... parameters)

Create and return a new moving average lag polynomial.

Parameters:

parameters - the moving average parameters of an ARIMA model.

Returns:

a new moving average lag polynomial.

autoRegressive

public static LagPolynomial autoRegressive(double... parameters)

Create and return a new autoregressive lag polynomial.

Parameters:

parameters - the autoregressive parameters of an ARIMA model.

Returns:

a new autoregressive lag polynomial.

times

public final LagPolynomial times(LagPolynomial other)

Multiply this polynomial by another lag polynomial and return the result in a new lag polynomial.

Parameters:

other - the polynomial to multiply this one with.

Returns:

the product of this polynomial with the given polynomial.

apply

public final double apply(TimeSeries timeSeries,
                          int index)

Apply this lag polynomial to a time series at the given index.

Parameters:

timeSeries - the time series containing the index to apply this lag polynomial to.

index - the index of the series to apply the lag polynomial at.

Returns:

the result of applying this lag polynomial to the given time series at the given index.

apply

public final double apply(TimeSeries timeSeries,
                          OffsetDateTime dateTime)

Apply this lag polynomial to a time series at the given date-time.

Parameters:

timeSeries - the time series containing the index to apply this lag polynomial to.

dateTime - the date-time of the series to apply the lag polynomial at.

Returns:

the result of applying this lag polynomial to the given time series at the given date-time.

solve

public double solve(TimeSeries timeSeries,
                    int index)

Apply this lag polynomial to the series at the given index, then solve for the value of the series at that index.

Parameters:

timeSeries - the time series containing the index to apply the lag polynomial to.

index - the index of the series to apply the lag polynomial at.

Returns:

the result of applying this lag polynomial to the time series at the given index, then solving for the value of the series at that index.

solve

public double solve(TimeSeries timeSeries,
                    OffsetDateTime dateTime)

Apply this lag polynomial to the series at the given date-time, then solve for the value of the series at that date-time.

Parameters:

timeSeries - the time series containing the date-time to apply the lag polynomial to.

dateTime - the date-time of the series to apply the lag polynomial at.

Returns:

the result of applying this lag polynomial to the time series at the given date-time, then solving for the value of the series at that date-time.

solve

public double solve(double[] timeSeries,
                    int index)

Apply this lag polynomial to the series at the given index, then solve for the value of the series at that index.

Parameters:

timeSeries - the time series containing the index to apply the lag polynomial to.

index - the index of the series to apply the lag polynomial at.

Returns:

the result of applying this lag polynomial to the time series at the given index, then solving for the value of the series at that index.

parameters

public final double[] parameters()

Return the parameters of this lag polynomial.

Returns:

the parameters of this lag polynomial.

inverseParams

public final double[] inverseParams()

Return the additive inverse of the parameters of this lag polynomial.

Returns:

the additive inverse of the parameters of this lag polynomial.

toString

public String toString()

Overrides:

toString in class Object

hashCode

public int hashCode()

Overrides:

hashCode in class Object

equals

public boolean equals(Object obj)

Overrides:

equals in class Object