Steps for feature columns specification.


List of steps that can be used to specify columns in the feature_spec interface.



  • step_numeric_column() to define numeric columns.

  • step_categorical_column_with_vocabulary_list() to define categorical columns.

  • step_categorical_column_with_hash_bucket() to define categorical columns where ids are set by hashing.

  • step_categorical_column_with_identity() to define categorical columns represented by integers in the range [0-num_buckets).

  • step_categorical_column_with_vocabulary_file() to define categorical columns when their vocabulary is available in a file.

  • step_indicator_column() to create indicator columns from categorical columns.

  • step_embedding_column() to create embeddings columns from categorical columns.

  • step_bucketized_column() to create bucketized columns from numeric columns.

  • step_crossed_column() to perform crosses of categorical columns.

  • step_shared_embeddings_column() to share embeddings between a list of categorical columns.

  • step_remove_column() to remove columns from the specification.

See Also

  • selectors for a list of selectors that can be used to specify variables.

    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_remove_column(), step_shared_embeddings_column()