Default forecasting models

Return the default set of forecasting models used by get_cv and get_fcast.

Usage

default_models()

Value

A named list of fable model specifications.

Details

The default models are naive, ETS, Theta, and ARIMA models. All models are fitted to log(observation + 1) to stabilise variance in count data. Forecasts are automatically transformed back to the original scale by fable. +1 is added to avoid taking the log of zero.

Additional models can be added by extending the returned list, for example: c(default_models(), list(CUSTOM = fable::ARIMA(observation))).

Author

Cyril Geismar

Examples

default_models()
#> $NAIVE
#> <RW model definition>
#> 
#> $ETS
#> <ETS model definition>
#> 
#> $THETA
#> <theta model definition>
#> 
#> $ARIMA
#> <ARIMA model definition>
#>