isomap.md
In recipes: Preprocessing and Feature Engineering Steps for Modeling

No ISOmap

Code
  print(im_rec)
Message

  -- Recipe ----------------------------------------------------------------------

  -- Inputs 
  Number of variables by role
  predictor: 3

  -- Training information 
  Training data contained 5 data points and no incomplete rows.

  -- Operations 
  * Isomap was not conducted for: x1, x2, x3 | Trained

ISOmap fails gracefully

Code
  prep(step_isomap(step_other(step_bs(step_bs(recipe(Sepal.Length ~ ., data = iris),
  Sepal.Width, deg_free = 1, degree = 1), Sepal.Length, deg_free = 1, degree = 1),
  Species, threshold = 1e-09), all_numeric_predictors(), num_terms = 1,
  neighbors = 1))
Message
Condition
  Error in `step_isomap()`:
  Caused by error in `prep()`:
  ! Failed to compute:
  Caused by error:
  ! TridiagEigen: eigen decomposition failed

check_name() is used

Code
  prep(rec, training = dat)
Message
Condition
  Error in `step_isomap()`:
  Caused by error in `bake()`:
  ! Name collision occurred. The following variable names already exist:
  * `Isomap1`

check_options() is used

Code
  prep(step_isomap(recipe(~mpg, data = mtcars), mpg, options = TRUE))
Condition
  Error in `step_isomap()`:
  Caused by error in `prep()`:
  ! `options` must be a list, not `TRUE`.

bake method errors when needed non-standard role columns are missing

Code
  bake(im_trained, new_data = dat2[, 1:2])
Condition
  Error in `step_isomap()`:
  ! The following required column is missing from `new_data`: x3.

empty printing

Code
  rec
Message

  -- Recipe ----------------------------------------------------------------------

  -- Inputs 
  Number of variables by role
  outcome:    1
  predictor: 10

  -- Operations 
  * Isomap approximation with: <none>

Code
  rec
Message

  -- Recipe ----------------------------------------------------------------------

  -- Inputs 
  Number of variables by role
  outcome:    1
  predictor: 10

  -- Training information 
  Training data contained 32 data points and no incomplete rows.

  -- Operations 
  * Isomap approximation with: <none> | Trained

keep_original_cols - can prep recipes with it missing

Code
  rec <- prep(rec)
Message
Condition
  Warning:
  `keep_original_cols` was added to `step_isomap()` after this recipe was created.
  i Regenerate your recipe to avoid this warning.

printing

Code
  print(rec)
Message

  -- Recipe ----------------------------------------------------------------------

  -- Inputs 
  Number of variables by role
  predictor: 3

  -- Operations 
  * Isomap approximation with: x1, x2, x3

Code
  prep(rec)
Message

  -- Recipe ----------------------------------------------------------------------

  -- Inputs 
  Number of variables by role
  predictor: 3

  -- Training information 
  Training data contained 5 data points and no incomplete rows.

  -- Operations 
  * Isomap approximation with: x1, x2, x3 | Trained

bad args

Code
  prep(step_isomap(recipe(~., data = mtcars), all_predictors(), num_terms = 2,
  neighbors = -1 / 3))
Condition
  Error in `step_isomap()`:
  Caused by error in `prep()`:
  ! `neighbors` must be a whole number, not the number -0.33.

Code
  prep(step_isomap(recipe(~., data = mtcars), all_predictors(), prefix = NULL))
Condition
  Error in `step_isomap()`:
  Caused by error in `prep()`:
  ! `prefix` must be a single string, not `NULL`.