tfdatasets (version 2.17.0)

step_remove_column: Creates a step that can remove columns

Description

Removes features of the feature specification.

Usage

step_remove_column(spec, ...)

Value

a FeatureSpec object.

Arguments

spec

A feature specification created with feature_spec().

...

Comma separated list of variable names to apply the step. selectors can also be used.

See Also

steps for a complete list of allowed steps.

Other Feature Spec Functions: dataset_use_spec(), feature_spec(), fit.FeatureSpec(), step_bucketized_column(), step_categorical_column_with_hash_bucket(), step_categorical_column_with_identity(), step_categorical_column_with_vocabulary_file(), step_categorical_column_with_vocabulary_list(), step_crossed_column(), step_embedding_column(), step_indicator_column(), step_numeric_column(), step_shared_embeddings_column(), steps

Examples

if (FALSE) {
library(tfdatasets)
data(hearts)
hearts <- tensor_slices_dataset(hearts) %>% dataset_batch(32)

# use the formula interface
spec <- feature_spec(hearts, target ~ age) %>%
  step_numeric_column(age, normalizer_fn = scaler_standard()) %>%
  step_bucketized_column(age, boundaries = c(20, 50)) %>%
  step_remove_column(age)

spec_fit <- fit(spec)
final_dataset <- hearts %>% dataset_use_spec(spec_fit)
}