tidymodels Novel levels found in column

Question

I am using tidymodels to create a Random Forrest prediction. I have test data that contains a new factor level not present in the training data which results in the error:

1: Novel levels found in column 'Siblings': '4'. The levels have been removed, and values have been coerced to 'NA'. 
2: There are new levels in a factor: NA 
> test_predict
Fehler: Objekt 'test_predict' nicht gefunden

I tried to include a step_novel and step_dummy on the "Siblings" column but this does not resolve the error. How should I deal with new factors not present in training data?

library(tidyverse)
library(tidymodels)

data <-
  data.frame(
    Survived = as.factor(c(0,1,1,1,0,0,0,0,0,1,1,1,0,0,0,0)),
    Siblings = as.factor(c(1,1,0,1,0,0,0,3,1,1,0,1,0,0,0,3)),
    Class = as.factor(c(0,1,0,1,0,1,0,0,0,1,0,1,0,1,0,0)),
    Embarked = as.factor(c("s","c","m","m","s","c","s","m","m","s","s","s","s","s","s","s")) 
  )

test <-
  data.frame(
    Siblings = as.factor(c(1,1,0,1,0,0,0,3,1,1,0,1,0,0,0,4)), #New factor level
    Class = as.factor(c(0,1,0,1,0,1,0,0,0,1,0,1,0,1,0,0)),
    Embarked = as.factor(c("s","c","m","m","s","c","s","m","m","s","s","s","s","s","s","s")) 
  )

#Model
rf_model <-
  rand_forest() %>%
  set_args(
    mtry = 3,
    trees = 1000,
    min_n = 15
  ) %>%
  set_engine("ranger", 
             importance = "impurity") %>%
  set_mode("classification")

#Recipe
data_recipe <- 
  recipe(Survived ~Siblings + Class + Embarked, data=data) %>%
  step_novel(Siblings) %>%
  step_dummy(Siblings)
#Workflow
rf_workflow <- 
  workflow() %>%
  add_recipe(data_recipe) %>%
  add_model(rf_model)

final_model <- fit(rf_workflow, data)
final_model

test_predict <- predict(final_model, test)
test_predict

Answer 1

If you notice in the documentation for step_novel(), it says:

When fitting a model that can deal with new factor levels, consider using workflows::add_recipe() with allow_novel_levels = TRUE set in hardhat::default_recipe_blueprint(). This will allow your model to handle new levels at prediction time, instead of throwing warnings or errors.

So you want to do that:

library(tidyverse)
library(tidymodels)
#> Registered S3 method overwritten by 'tune':
#>   method                   from   
#>   required_pkgs.model_spec parsnip

data <-
  data.frame(
    Survived = as.factor(c(0,1,1,1,0,0,0,0,0,1,1,1,0,0,0,0)),
    Siblings = as.factor(c(1,1,0,1,0,0,0,3,1,1,0,1,0,0,0,3)),
    Class = as.factor(c(0,1,0,1,0,1,0,0,0,1,0,1,0,1,0,0)),
    Embarked = as.factor(c("s","c","m","m","s","c","s","m","m","s","s","s","s","s","s","s")) 
  )

test <-
  data.frame(
    Siblings = as.factor(c(1,1,0,1,0,0,0,3,1,1,0,1,0,0,0,4)), #New factor level
    Class = as.factor(c(0,1,0,1,0,1,0,0,0,1,0,1,0,1,0,0)),
    Embarked = as.factor(c("s","c","m","m","s","c","s","m","m","s","s","s","s","s","s","s")) 
  )

#Model
rf_model <-
  rand_forest() %>%
  set_args(
    mtry = 3,
    trees = 1000,
    min_n = 15
  ) %>%
  set_engine("ranger", 
             importance = "impurity") %>%
  set_mode("classification")

#Recipe
data_recipe <- 
  recipe(Survived ~Siblings + Class + Embarked, data=data) %>%
  step_novel(Siblings) %>%
  step_dummy(Siblings)

#Workflow
rf_workflow <- 
  workflow() %>%
  add_recipe(data_recipe, 
             blueprint = hardhat::default_recipe_blueprint(allow_novel_levels = TRUE)) %>%
  add_model(rf_model)

final_model <- fit(rf_workflow, data)
final_model
#> ══ Workflow [trained] ══════════════════════════════════════════════════════════
#> Preprocessor: Recipe
#> Model: rand_forest()
#> 
#> ── Preprocessor ────────────────────────────────────────────────────────────────
#> 2 Recipe Steps
#> 
#> • step_novel()
#> • step_dummy()
#> 
#> ── Model ───────────────────────────────────────────────────────────────────────
#> Ranger result
#> 
#> Call:
#>  ranger::ranger(x = maybe_data_frame(x), y = y, mtry = min_cols(~3,      x), num.trees = ~1000, min.node.size = min_rows(~15, x),      importance = ~"impurity", num.threads = 1, verbose = FALSE,      seed = sample.int(10^5, 1), probability = TRUE) 
#> 
#> Type:                             Probability estimation 
#> Number of trees:                  1000 
#> Sample size:                      16 
#> Number of independent variables:  5 
#> Mtry:                             3 
#> Target node size:                 15 
#> Variable importance mode:         impurity 
#> Splitrule:                        gini 
#> OOB prediction error (Brier s.):  0.254242

test_predict <- predict(final_model, test)
test_predict
#> # A tibble: 16 x 1
#>    .pred_class
#>    <fct>      
#>  1 0          
#>  2 1          
#>  3 0          
#>  4 1          
#>  5 0          
#>  6 0          
#>  7 0          
#>  8 0          
#>  9 0          
#> 10 1          
#> 11 0          
#> 12 1          
#> 13 0          
#> 14 0          
#> 15 0          
#> 16 0

^{Created on 2021-07-09 by the reprex package (v2.0.0)}

The workflows functions are very strict about factor levels and other aspects of the new data, ensuring that they match up with the training data.

Answer 2

To answer my own question:

We need to apply step_novel followed by step_unknown. As far as I understand from the documentation step_novel labels any new factors occurring in the data with "new". This can be used to easily identify such factors when the data is inspected after applying the recipe. step_unknown removes any such factors from the data and converts the values to NA when the model is applied:

data_recipe <- 
   recipe(Survived ~Siblings + Class + Embarked, data=data) %>%
   step_novel(Siblings) %>%
   step_unknown(Siblings)