Modeltime best workflow from a set of models

get best workflows generated from the modeltime_wfs_fit() function output.

modeltime_wfs_bestmodel(
  .wfs_results,
  .model = NULL,
  .metric = "rmse",
  .minimize = TRUE
)

Arguments

.wfs_results	a tibble generated from the modeltime_wfs_fit() function.
.model	string or number, It can be supplied as follows: “top n,” “Top n” or “tOp n”, where n is the number of best models to select; n, where n is the number of best models to select; name of the workflow or workflows to select.
.metric	metric to get best model from ('mae', 'mape','mase','smape','rmse','rsq')
.minimize	a boolean indicating whether to minimize (TRUE) or maximize (FALSE) the metric.

Value

a tibble containing the best model based on the selected metric.

Details

the best model is selected based on a specific metric ('mae', 'mape','mase','smape','rmse','rsq'). The default is to minimize the metric. However, if the model is being selected based on rsq minimize should be FALSE.

Examples

library(dplyr)
library(earth)
data <- sknifedatar::data_avellaneda %>% mutate(date=as.Date(date)) %>% filter(date<'2012-06-01')

recipe_date <- recipes::recipe(value ~ ., data = data) %>% 
  recipes::step_date(date, features = c('dow','doy','week','month','year')) 

mars <- parsnip::mars(mode = 'regression') %>%
  parsnip::set_engine('earth')

wfsets <- workflowsets::workflow_set(
  preproc = list(
    R_date = recipe_date),
  models  = list(M_mars = mars),
  cross   = TRUE)

wffits <- sknifedatar::modeltime_wfs_fit(.wfsets = wfsets, 
                                         .split_prop = 0.8, 
                                         .serie=data)
#> 
#> ── MODEL: R_date_M_mars 
#> ✓ Training finished OK.
#> 
#> 
#> ── 1 models fitted ♥ ───────────────────────────────────────────────────────────
#> 
#> ── 0 models deleted x ──
#> 

sknifedatar::modeltime_wfs_bestmodel(.wfs_results = wffits,
                                     .metric='rsq',
                                     .minimize = FALSE)
#> # A tibble: 1 x 4
#>   .model_id      rank .model_desc .fit_model          
#>   <chr>         <int> <chr>       <list>              
#> 1 R_date_M_mars     1 EARTH       <model_time [1 × 3]>