Forecaster
TSB
Teunter-Syntetos-Babai method for forecasting intermittent time series.
Implements the method proposed by Teunter, Syntetos, Babai in [1].
The TSB method is a modification of Croston’s method to handle intermittent time series with very sporadic demand, providing a more accurate estimation of the risk of obsolescence. Instead of forecasting the demand interval, TSB forecasts demand probability which is updated every period.
TSB will predict a constant value for all future times, so it essentially provides another notion for the average value of a time series.
Schnellstart
python
from sktime.forecasting.tsb import TSB
estimator = TSB(alpha=0.1, beta=0.1)Tags
Fähigkeiten
- Kategoriale Merkmale
- In-Sample-Vorhersagen: Unterstützt
- In-Sample-Prognoseintervalle: Unterstützt
- Exogene Variablen
- Prognoseintervalle: Nicht unterstützt
- Fehlende Werte: Nicht unterstützt
- Multivariat: Nicht unterstützt
Eigenschaften
- Prognosehorizont beim Fit nötigrequires-fh-in-fit
- Nein
- X und y brauchen denselben IndexX-y-must-have-same-index
- Ja
Parameter(2)
- alphafloat, default = 0.1
- Smoothing parameter for demand size (d)
- betafloat, default = 0.1
- Smoothing parameter for demand occurrence probability (p)
Beispiele
>>> from sktime.forecasting.tsb import TSB
>>> from sktime.datasets import load_PBS_dataset
>>> y = load_PBS_dataset ()
>>> forecaster = TSB (alpha = 0.4, beta = 0.05)
>>> forecaster. fit (y) TSB(
... )
>>> y_pred = forecaster. predict (fh = [1, 2, 3 ])Referenzen
[1]
Teunter, R.H., Syntetos, A.A. and Zied Babai, M. (2011)
Intermittent Demand: Linking Forecasting to Inventory Obsolescence. European Journal of Operational Research, 214, 606-615.