tidymodels: How do I get list items returned by original models?

Hello world! I am starting to like tidymodels. It gives me the tidyverse friendly coding styles.

I have a short question.

How do I get the list items of models called by parsnip package? A quick example as followed

library(tidymodels)
# for instance, i want to get "importance" list items by returned randomForest::randomForest
original.model <- randomForest::randomForest(Species ~ ., data=iris, ntree=1000, mtry=2, importance=TRUE)

original.model$importance
importance(original.model)
varImpPlot(original.model)
#              setosa versicolor
# Sepal.Length 0.0333    0.02056
# Sepal.Width  0.0078    0.00283
# Petal.Length 0.3289    0.30913
# Petal.Width  0.3316    0.30632
#              virginica
# Sepal.Length   0.04028
# Sepal.Width    0.00946
# Petal.Length   0.29844
# Petal.Width    0.26959

# from all the possible list items
names(original.model)
#>  [1] "call"            "type"            "predicted"      
#>  [4] "err.rate"        "confusion"       "votes"          
#>  [7] "oob.times"       "classes"         "importance"     
#> [10] "importanceSD"    "localImportance" "proximity"      
#> [13] "ntree"           "mtry"            "forest"         
#> [16] "y"               "test"            "inbag"          
#> [19] "terms"   

parsnip.model <- rand_forest(mode='classification', mtry=2, trees=1000) %>%
  set_engine('randomForest', importance=TRUE) %>%
  fit(Species ~ ., data=iris)

# how do I get the list items as above?
names(parsnip.model)
#> [1] "lvl"     "spec"    "fit"     "preproc"

The underlying model object is in the fit slot.

2 Likes

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.