The claim
Depending on the value of its complex smoothing parameter, CES is a stationary or a non-stationary model.
Proof: E3 · Uses: C17, C18, C19, C20, C23, C24, C25, C32, C21, C22, C33 · Extends: Hyndman et al. (2008) · Addresses: B3 · Addresses: B2 · Addresses: B4 · Addresses: B5 · Verification: not checked
Claims it rests on
CES is stationary, with both eigenvalues of its transition matrix inside the unit circle, when , and .
Proof: E3 · Cited: Hyndman et al. (2008), p. 38, own earlier work · Verification: not checked · Used by: C1, C24
When , both eigenvalues of the CES transition matrix are real and one of them is greater than one.
When , CES produces a non-stationary trajectory.
Proof: E3 · Uses: C18 · Verification: not checked · Used by: C1
When , CES is equivalent to ETS(A,N,N).
CES is required to be stable, or at least forecastable.
Cited: Ord et al. (1997), own earlier work · Cited: Hyndman et al. (2008), own earlier work · Used by: C1
CES is not required to be stationary.
Cited: Ord et al. (1997), own earlier work · Cited: Hyndman et al. (2008), own earlier work · Used by: C1
CES is stable, with all eigenvalues of its discount matrix inside the unit circle, when , and .
Proof: E4 · Cited: Hyndman et al. (2008), pp. 172-173, own earlier work · Verification: not checked · Used by: C1, C24
The stable region of CES contains both stationary and non-stationary values of the complex smoothing parameter.
Proof: E4 · Uses: C17, C23 · Verification: not checked · Used by: C1
The CES point forecast follows one of five trajectories set by the complex smoothing parameter: a flat line, exponential growth, exponential decline, harmonic oscillation converging to zero, or diverging harmonic oscillation.
Proof: E3 · Uses: C26, C27, C28, C29, C30 · Explains: C36 · Explains: C37 · Verification: not checked · Used by: C1
When , the CES point forecast grows exponentially.
When , the CES point forecast declines exponentially.
When , the CES point forecast is a harmonic oscillation converging to zero.
When , the CES point forecast is a diverging harmonic oscillation.
On M3 series 1664, which graphical analysis and ADF test and KPSS test indicate is stationary, CES estimates the stationary complex smoothing parameter .
On the trended M3 series 2721, CES estimates the non-stationary complex smoothing parameter .
Evidence
Analysis of the eigenvalues of the CES transition matrix: the stationarity condition, the non-stationary case, the special case equivalent to ETS(A,N,N), the five forecast trajectories and the connection with simple exponential smoothing.
Foundation: Linear algebra: eigenvalues of the transition matrix of a linear state-space model · Assumes: C61 · Proof: A9
Derivation of the stability region of CES from the invertibility conditions of its underlying ARMA(2,2) model.
Foundation: Invertibility conditions of ARMA(2,2) models; algebra of inequalities · Assumes: C61 · Proof: A9
CES fitted with ces() from smooth to M3 series 1664 and 2721, each split into in-sample and holdout parts; stationarity of each series assessed by graphical analysis and ADF test and KPSS test; point forecasts and 95% prediction intervals.
Setup: studied CES · cases M3 series 1664 and 2721 · procedure ADF test, KPSS test · lens complex smoothing parameter
- R1 complex smoothing parameter = 1 (part: real part alpha_0, series: M3 series 1664)
- R2 complex smoothing parameter = 0.9988 (part: imaginary part alpha_1, series: M3 series 1664)
- R3 complex smoothing parameter = 1.481 (part: real part alpha_0, series: M3 series 2721)
- R4 complex smoothing parameter = 1.0035 (part: imaginary part alpha_1, series: M3 series 2721)
Assumptions
Definitions
Methods 2
D1 CES: Complex exponential smoothing: a forecasting method that combines each past observation and its forecast error into the complex variable and forecasts with an exponentially weighted sum of these values, or recursively where is the complex smoothing parameter and carries the information on past forecast errors. Assumes C60, C62, C63, C64.
D19 ETS(A,N,N): The ETS model with additive error, no trend and no seasonality, fitted without model selection. Source: Hyndman et al. (2008).
Data 1
D42 M3 series 1664 and 2721: Two monthly series of the M3 competition: series 1664, which graphical analysis and the ADF and KPSS tests indicate is stationary, and the trended series 2721; each split into in-sample and holdout parts.
Variables and measures 1
D2 complex smoothing parameter: The smoothing parameter of CES, the complex number ; its real part and imaginary part are the two real parameters estimated, and they set how much of the past observations and of the past forecast errors is carried into the forecast.
Study design 2
D40 ADF test: Augmented Dickey-Fuller test of the null hypothesis that a time series has a unit root. Source: Dickey and Fuller (1979).
D41 KPSS test: Kwiatkowski-Phillips-Schmidt-Shin test of the null hypothesis that a time series is stationary. Source: Kwiatkowski et al. (1992).
Context
- Extends Hyndman et al. (2008) · own earlier work · from C1
- Exponential smoothing in a state space framework (ETS) gives 30 models with no, additive or multiplicative error, trend and seasonal components, among which a model is selected by an information criterion.
- Addresses B3 · own earlier work · from C1
- Identifying the presence and type of trend in exponential smoothing often leads to misidentification and poorly performing long-term forecasts.
- Addresses B2 · from C1
- The diverse trend alternatives of exponential smoothing in a state space framework make the selection of a model difficult.
- Addresses B4 · from C1
- Selecting the most appropriate exponential smoothing model is a challenging task.
- Addresses B5 · own earlier work · from C1
- Even when the components of exponential smoothing are identified correctly, the estimated smoothing parameters and the weights they imply can give poor forecasts.
References
- Dickey, D. A., Fuller, W. A., 1979. Distribution of the Estimators for Autoregressive Time Series with a Unit Root. Journal of the American Statistical Association 74, 427–431. https://doi.org/10.1080/01621459.1979.10482531 Used by: D40
- Hyndman, R., Koehler, A., Ord, K., Snyder, R., 2008. Forecasting with Exponential Smoothing. Springer Series in Statistics. https://doi.org/10.1007/978-3-540-71918-2 Used by: Context, C17, C21, C22, C23, D19
- Kwiatkowski, D., Phillips, P. C. B., Schmidt, P., Shin, Y., 1992. Testing the null hypothesis of stationarity against the alternative of a unit root. Journal of Econometrics 54, 159–178. https://doi.org/10.1016/0304-4076(92)90104-Y Used by: D41
- Makridakis, S., Hibon, M., 2000. The M3-Competition: results, conclusions and implications. International Journal of Forecasting 16, 451–476. https://doi.org/10.1016/S0169-2070(00)00057-1 Used by: A2
- Ord, J. K., Koehler, A. B., Snyder, R. D., 1997. Estimation and Prediction for a Class of Dynamic Nonlinear Statistical Models. Journal of the American Statistical Association 92, 1621–1629. https://doi.org/10.1080/01621459.1997.10473684 Used by: C21, C22
- R Core Team (2021). R: A language and environment for statistical computing. R Foundation for Statistical Computing. https://www.R-project.org/ Used by: A8
- Svetunkov, I., Kourentzes, N., Ord, J. K., 2022. Complex exponential smoothing. Naval Research Logistics (NRL) 69, 1108–1123. https://doi.org/10.1002/nav.22074 Used by: A9
- https://github.com/config-i1/smooth Used by: A4
Artefacts
- A2 · data · Makridakis and Hibon (2000) M3 competition data
- A4 · software · https://github.com/config-i1/smooth smooth package for R (ces, auto.ces, es, auto.ssarima) smooth_3.1.2.tar.gz
- A8 · software · R Core Team (2021). R: A language and environment for statistical computing. R Foundation for Statistical Computing. https://www.R-project.org/ (2021) R, the environment for statistical computing in which the experiments were run www.R-project.org
- A9 · manuscript · Svetunkov et al. (2022) The published article (open access, CC BY). nav.22074
Score
Scores are computed by rules from the claimset file (EntryLib rubric 0.7): they check the record, not the science, and may change as the rubric and reviews develop. The rubric.
C1: 65, Partly supported
Novelty: extends earlier work, by the authors
- −25 Evidence Rests on argument only; an unchecked proof; evidence named without the observations the claim reports; an argument from other claims; cited work only. To remove it: Support the claim with evidence the claimset records, citing the result entries or table rows it reports, or have its proof checked; for cited work, read the work in full and quote the passage the claim rests on from a captured copy (claimset extension review).
- −10 Context No scope: where the claim holds is not stated. To remove it: Give the claim a scope: population, setting, period, data, conditions, exclusions.
C17: 80, Well supported
Novelty: new, by the authors
- −20 Evidence Rests on argument only; an unchecked proof; evidence named without the observations the claim reports; an argument from other claims; cited work only. To remove it: Support the claim with evidence the claimset records, citing the result entries or table rows it reports, or have its proof checked; for cited work, read the work in full and quote the passage the claim rests on from a captured copy (claimset extension review).
C18: 75, Partly supported
Novelty: new, by the authors
- −25 Evidence Rests on argument only; an unchecked proof; evidence named without the observations the claim reports; an argument from other claims; cited work only. To remove it: Support the claim with evidence the claimset records, citing the result entries or table rows it reports, or have its proof checked; for cited work, read the work in full and quote the passage the claim rests on from a captured copy (claimset extension review).
C19: 75, Partly supported
Novelty: new, by the authors
- −25 Evidence Rests on argument only; an unchecked proof; evidence named without the observations the claim reports; an argument from other claims; cited work only. To remove it: Support the claim with evidence the claimset records, citing the result entries or table rows it reports, or have its proof checked; for cited work, read the work in full and quote the passage the claim rests on from a captured copy (claimset extension review).
C20: 75, Partly supported
Novelty: new, by the authors
- −25 Evidence Rests on argument only; an unchecked proof; evidence named without the observations the claim reports; an argument from other claims; cited work only. To remove it: Support the claim with evidence the claimset records, citing the result entries or table rows it reports, or have its proof checked; for cited work, read the work in full and quote the passage the claim rests on from a captured copy (claimset extension review).
C21: 70, Partly supported
Novelty: new, by the authors
- −20 Evidence Rests on argument only; an unchecked proof; evidence named without the observations the claim reports; an argument from other claims; cited work only. To remove it: Support the claim with evidence the claimset records, citing the result entries or table rows it reports, or have its proof checked; for cited work, read the work in full and quote the passage the claim rests on from a captured copy (claimset extension review).
- −10 Context No scope: where the claim holds is not stated. To remove it: Give the claim a scope: population, setting, period, data, conditions, exclusions.
C22: 70, Partly supported
Novelty: new, by the authors
- −20 Evidence Rests on argument only; an unchecked proof; evidence named without the observations the claim reports; an argument from other claims; cited work only. To remove it: Support the claim with evidence the claimset records, citing the result entries or table rows it reports, or have its proof checked; for cited work, read the work in full and quote the passage the claim rests on from a captured copy (claimset extension review).
- −10 Context No scope: where the claim holds is not stated. To remove it: Give the claim a scope: population, setting, period, data, conditions, exclusions.
C23: 80, Well supported
Novelty: new, by the authors
- −20 Evidence Rests on argument only; an unchecked proof; evidence named without the observations the claim reports; an argument from other claims; cited work only. To remove it: Support the claim with evidence the claimset records, citing the result entries or table rows it reports, or have its proof checked; for cited work, read the work in full and quote the passage the claim rests on from a captured copy (claimset extension review).
C24: 75, Partly supported
Novelty: new, by the authors
- −25 Evidence Rests on argument only; an unchecked proof; evidence named without the observations the claim reports; an argument from other claims; cited work only. To remove it: Support the claim with evidence the claimset records, citing the result entries or table rows it reports, or have its proof checked; for cited work, read the work in full and quote the passage the claim rests on from a captured copy (claimset extension review).
C25: 75, Partly supported
Novelty: new, by the authors
- −25 Evidence Rests on argument only; an unchecked proof; evidence named without the observations the claim reports; an argument from other claims; cited work only. To remove it: Support the claim with evidence the claimset records, citing the result entries or table rows it reports, or have its proof checked; for cited work, read the work in full and quote the passage the claim rests on from a captured copy (claimset extension review).
C26: 75, Partly supported
Novelty: new, by the authors
- −25 Evidence Rests on argument only; an unchecked proof; evidence named without the observations the claim reports; an argument from other claims; cited work only. To remove it: Support the claim with evidence the claimset records, citing the result entries or table rows it reports, or have its proof checked; for cited work, read the work in full and quote the passage the claim rests on from a captured copy (claimset extension review).
C27: 75, Partly supported
Novelty: new, by the authors
- −25 Evidence Rests on argument only; an unchecked proof; evidence named without the observations the claim reports; an argument from other claims; cited work only. To remove it: Support the claim with evidence the claimset records, citing the result entries or table rows it reports, or have its proof checked; for cited work, read the work in full and quote the passage the claim rests on from a captured copy (claimset extension review).
C28: 75, Partly supported
Novelty: new, by the authors
- −25 Evidence Rests on argument only; an unchecked proof; evidence named without the observations the claim reports; an argument from other claims; cited work only. To remove it: Support the claim with evidence the claimset records, citing the result entries or table rows it reports, or have its proof checked; for cited work, read the work in full and quote the passage the claim rests on from a captured copy (claimset extension review).
C29: 75, Partly supported
Novelty: new, by the authors
- −25 Evidence Rests on argument only; an unchecked proof; evidence named without the observations the claim reports; an argument from other claims; cited work only. To remove it: Support the claim with evidence the claimset records, citing the result entries or table rows it reports, or have its proof checked; for cited work, read the work in full and quote the passage the claim rests on from a captured copy (claimset extension review).
C30: 75, Partly supported
Novelty: new, by the authors
- −25 Evidence Rests on argument only; an unchecked proof; evidence named without the observations the claim reports; an argument from other claims; cited work only. To remove it: Support the claim with evidence the claimset records, citing the result entries or table rows it reports, or have its proof checked; for cited work, read the work in full and quote the passage the claim rests on from a captured copy (claimset extension review).
C32: 90, Well supported
Novelty: new, by the authors
- −10 Context No scope: where the claim holds is not stated. To remove it: Give the claim a scope: population, setting, period, data, conditions, exclusions.
C33: 90, Well supported
Novelty: new, by the authors
- −10 Context No scope: where the claim holds is not stated. To remove it: Give the claim a scope: population, setting, period, data, conditions, exclusions.