ETS

ETS (Error, Trend, Seasonality) is a forecasting model that decomposes a time series into error, trend, and seasonal components.

Error

Try Multiplicative (M) Try Additive (A)

Trend

Try Multiplicative (M) Try Additive (A) Try None (N)

Damped trend

Try regular trend Try damped trend (d)

Seasonality

Try Multiplicative (M) Try Additive (A) Try None (N)

Season length

Number of time steps in a season.

Random state

Using a fixed random seed allows for reproducible result.

Unstable models

Include unstable models.

Several parameter combinations are known to be unstable (AMN, AMA, AMdN, AMdA, AMM, AMdM, MMA, MMdA, ANM, AAM, AAdM).

Unstable combinations are likely to fail when training and scoring.

Input data must be strictly positive when using multiplicative error, trend or seasonal components.

Damped trend requires at least one Trend parameter to be different from None.