Better description/information on Exception: no model able to be fitted

[renanklehm]

I'm trying to fit several models in a group of time series, using the following code:

fcst = StatsForecast(df=Y_train_df, models=[ AutoARIMA(season_length=12), AutoCES(season_length=12), SeasonalNaive(season_length=12), HistoricAverage(), RandomWalkWithDrift(), ], freq='M', n_jobs=-1)

But sometimes I get the exception "no model able to be fitted", which just informs me that there is something wrong. It would be nice if the exception were more informative, describing which unique_id and model failed to fit.

[mergenthaler]

Thanks for creating the issue @renanklehm, this has been requested before and we are working on it.

Meanwhile, you might find the fallback_model argument useful. You can check the documentation here.

Fit the models using Naive as a fallback model: fcst = StatsForecast(df=Y_train_df, models=[ AutoARIMA(season_length=12), AutoCES(season_length=12), SeasonalNaive(season_length=12), HistoricAverage(), RandomWalkWithDrift(), Naive() ], fallback_model=Naive(), freq='M', n_jobs=-1)