How to rpart.plot for testing data set ?

Machine Learning and Modeling
,
1

Hello All,
I am building decision tree with normal steps:

1. Split the data:

split <- policy %>%
  initial_split(prop = 0.80, strata = Lapsed)
train <- training(policy_split)
test <- testing(policy_split)
folds <- vfold_cv(train, strata = Lapsed)

2. Build Decision Tree model

set.seed(123)
tree_spec <- decision_tree() %>%
  set_engine("rpart") %>% 
  set_mode("classification")

tree_wf <- workflow() %>%
  add_recipe(pol_rec) %>%
  add_model(tree_spec)

set.seed(123)

tree_fit <- fit_resamples(
    tree_wf,
    resamples = folds,
    metrics = metric_set(accuracy,roc_auc,sens, spec),
    control = control_resamples(save_pred = TRUE))

3. Final model with last_fit()

fn_metrics <- metric_set(roc_auc, accuracy, sens, spec)
# Decision Tree
tree_final <- last_fit(
  tree_wf,
  split = split,
  metrics = fn_metrics
)

My question is "How I can plot the final Decision Tree for testing data set ?" instead of training data as below code:

tree_final %>%
  extract_fit_engine() %>%
  rpart.plot::rpart.plot(type = 4, extra = 2,
                         roundint = FALSE)

Thank you for your help.
Tam Pham

1 Like
2

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

If you have a query related to it or one of the replies, start a new topic and refer back with a link.