Forecasting of multiple models over multiple time series

allows forecasting on multiple time series from multiple fitted models.

modeltime_multiforecast(models_table, .h = NULL, .prop = NULL)

Arguments

models_table	'table_time' tibble generated with the modeltime_multifit() function.
.h	prediction horizon of the modeltime_forecast() function.
.prop	time series split partition ratio. If "h" is specified, this function predicts on the testing partition.

Value

'models_table' tibble with a new column called 'nested_forecast' where the predictions are stored.

Details

this function takes the 'table_time' object generated with the modeltime_multifit() function, the modeltime_forecast() from the package 'modeltime' is applied to each model for each series.

Examples

# Data
data_serie <- sknifedatar::table_time
                                     
# Forecast
sknifedatar::modeltime_multiforecast(data_serie$table_time, .prop=0.8)
#> # A tibble: 2 x 7
#>   sector  nested_column  m_ets m_nnetar nested_model calibration nested_forecast
#>   <chr>   <list>         <lis> <list>   <list>       <list>      <list>         
#> 1 Comerc… <tibble [49 ×… <fit… <fit[+]> <model_time… <model_tim… <tibble [69 × …
#> 2 Ensena… <tibble [49 ×… <fit… <fit[+]> <model_time… <model_tim… <tibble [69 × …