Fits one or more tidymodels workflow objects to nested time series data using the following process:
Models are iteratively fit to training splits.
Accuracy is calculated on testing splits and is logged.
Accuracy results can be retrieved with extract_nested_test_accuracy()
Any model that returns an error is logged.
Error logs can be retrieved with extract_nested_error_report()
Forecast is predicted on testing splits and is logged.
Forecast results can be retrieved with extract_nested_test_forecast()
modeltime_nested_fit(
nested_data,
...,
model_list = NULL,
metric_set = default_forecast_accuracy_metric_set(),
conf_interval = 0.95,
conf_method = "conformal_default",
control = control_nested_fit()
)Nested time series data
Tidymodels workflow objects that will be fit to the nested time series data.
Optionally, a list() of Tidymodels workflow objects can be provided
A yardstick::metric_set() that is used to summarize one or more
forecast accuracy (regression) metrics.
An estimated confidence interval based on the calibration data. This is designed to estimate future confidence from out-of-sample prediction error.
Algorithm used to produce confidence intervals. All CI's are Conformal Predictions. Choose one of:
conformal_default: Uses qnorm() to compute quantiles from out-of-sample (test set) residuals.
conformal_split: Uses the split method split conformal inference method described by Lei et al (2018)
Used to control verbosity and parallel processing. See control_nested_fit().
Use extend_timeseries(), nest_timeseries(), and split_nested_timeseries() for preparing
data for Nested Forecasting. The structure must be a nested data frame, which is suppplied in
modeltime_nested_fit(nested_data).
Models must be in the form of tidymodels workflow objects. The models can be provided in two ways:
Using ... (dots): The workflow objects can be provided as dots.
Using model_list parameter: You can supply one or more workflow objects that are wrapped in a list().
A control object can be provided during fitting to adjust the verbosity and parallel processing.
See control_nested_fit().