kieker.tools.opad.timeseries.forecast.croston

Class CrostonForecaster

java.lang.Object

kieker.tools.opad.timeseries.forecast.AbstractForecaster<java.lang.Double>

kieker.tools.opad.timeseries.forecast.AbstractRForecaster

kieker.tools.opad.timeseries.forecast.croston.CrostonForecaster

All Implemented Interfaces:

IForecaster<java.lang.Double>

public class CrostonForecaster
extends AbstractRForecaster

Intermittent Demand Forecasting. This is one of the forecasters used in the research paper on Self-adaptive workload classification and forecasting for proactive resource provisioning (http://dx.doi.org/10.1002/cpe.3224), authored by Herbst et al.

Since:

1.10

Author:

Nikolas Herbst Decomposition of the time series that contains zero values into two separate sequences: a non-zero valued time series and a second that contains the time intervals of zero values. Independent forecast using SES and combination of the two independent forecasts. No confidence intervals are computed due to no consistent underlying stochastic model.

Field Summary

Fields inherited from class kieker.tools.opad.timeseries.forecast.AbstractRForecaster

MIN_TS_SIZE_DEFAULT

Constructor Summary

Constructors

Constructor and Description
CrostonForecaster(ITimeSeries<java.lang.Double> historyTimeseries)
CrostonForecaster(ITimeSeries<java.lang.Double> historyTimeseries, int confidenceLevel)

Method Summary

Methods

Modifier and Type	Method and Description
protected java.lang.String[]	getForecastFuncParams() Returns additional parameters to be appended to the call of the R forecast function.
protected java.lang.String[]	getModelFuncParams() Returns additional parameters to be appended to the call of the R forecast model.
protected boolean	supportsConfidence()

Methods inherited from class kieker.tools.opad.timeseries.forecast.AbstractRForecaster

createNaNForecast, forecast, forecastOperationOnResult, lowerOperationOnResult, removeNullValues, satisfiesInputTSRequirements, upperOperationOnResult

Methods inherited from class kieker.tools.opad.timeseries.forecast.AbstractForecaster

getConfidenceLevel, getTsOriginal, prepareForecastTS

Methods inherited from class java.lang.Object

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

Constructor Detail

CrostonForecaster

public CrostonForecaster(ITimeSeries<java.lang.Double> historyTimeseries)

Parameters:

historyTimeseries - timeseries used by forecating algo

CrostonForecaster

public CrostonForecaster(ITimeSeries<java.lang.Double> historyTimeseries,
                 int confidenceLevel)

Parameters:

historyTimeseries - timeseries used by forecating algo

confidenceLevel - value of confidence

Method Detail

getModelFuncParams

protected java.lang.String[] getModelFuncParams()

Description copied from class: AbstractRForecaster

Returns additional parameters to be appended to the call of the R forecast model.

Specified by:

getModelFuncParams in class AbstractRForecaster

Returns:

the parameters or null if none

getForecastFuncParams

protected java.lang.String[] getForecastFuncParams()

Description copied from class: AbstractRForecaster

Returns additional parameters to be appended to the call of the R forecast function.

Specified by:

getForecastFuncParams in class AbstractRForecaster

Returns:

the parameters or null if none

supportsConfidence

protected boolean supportsConfidence()

Overrides:

supportsConfidence in class AbstractForecaster<java.lang.Double>