Closed pikawangwang closed 2 months ago
Hi @pikawangwang,
You can find this information in the docstring of historical_forecasts()
: "This method repeatedly builds a training set: either expanding from the beginning of series or moving with a fixed length train_length
. It trains the model on the training set, emits a forecast of length equal to forecast_horizon, and then moves the end of the training set forward by stride
time steps.
By default, this method will return one (or a sequence of) single time series made up of the last point of each historical forecast. This time series will thus have a frequency of series.freq * stride. If last_points_only
is set to False, it will instead return one (or a sequence of) list of the historical forecasts series."
This method does not compute any loss/metric, but generate forecasts as if the model was fitted and called "from the past", to mimic historic/real-world usage of the model. Which can then be used to assess the model performance.
Does it make things clearer?
@madtoinou Thanks for your explanation, I understand now😀
I am a student trying to do time series forecasting.
When I am trying to using the function historical forecast with training set, I am confused about how the training loss is generated.
Is there a default fold of cross validaion, or the loss is generated by all of the training set?
Thank you!