tf.feature_column.crossed_column

TensorFlow 1 version View source on GitHub

Returns a column for performing crosses of categorical features.

Crossed features will be hashed according to hash_bucket_size. Conceptually, the transformation can be thought of as: Hash(cartesian product of features) % hash_bucket_size

For example, if the input features are:

  • SparseTensor referred by first key:
shape = [2, 2]
{
    [0, 0]: "a"
    [1, 0]: "b"
    [1, 1]: "c"
}
  • SparseTensor referred by second key:
shape = [2, 1]
{
    [0, 0]: "d"
    [1, 0]: "e"
}

then crossed feature will look like:

 shape = [2, 2]
{
    [0, 0]: Hash64("d", Hash64("a")) % hash_bucket_size
    [1, 0]: Hash64("e", Hash64("b")) % hash_bucket_size
    [1, 1]: Hash64("e", Hash64("c")) % hash_bucket_size
}

Here is an example to create a linear model with crosses of string features:

keywords_x_doc_terms = crossed_column(['keywords', 'doc_terms'], 50K)
columns = [keywords_x_doc_terms, ...]
features = tf.io.parse_example(..., features=make_parse_example_spec(columns))
linear_prediction = linear_model(features, columns)

You could also use vocabulary lookup before crossing:

keywords = categorical_column_with_vocabulary_file(
    'keywords', '/path/to/vocabulary/file', vocabulary_size=1K)
keywords_x_doc_terms = crossed_column([keywords, 'doc_terms'], 50K)
columns = [keywords_x_doc_terms, ...]
features = tf.io.parse_example(..., features=make_parse_example_spec(columns))
linear_prediction = linear_model(features, columns)

If an input feature is of numeric type, you can use categorical_column_with_identity, or bucketized_column, as in the example:

# vertical_id is an integer categorical feature.
vertical_id = categorical_column_with_identity('vertical_id', 10K)
price = numeric_column('price')
# bucketized_column converts numerical feature to a categorical one.
bucketized_price = bucketized_column(price, boundaries=[...])
vertical_id_x_price = crossed_column([vertical_id, bucketized_price], 50K)
columns = [vertical_id_x_price, ...]
features = tf.io.parse_example(..., features=make_parse_example_spec(columns))
linear_prediction = linear_model(features, columns)

To use crossed column in DNN model, you need to add it in an embedding column as in this example:

vertical_id_x_price = crossed_column([vertical_id, bucketized_price], 50K)
vertical_id_x_price_embedded = embedding_column(vertical_id_x_price, 10)
dense_tensor = input_layer(features, [vertical_id_x_price_embedded, ...])

keys An iterable identifying the features to be crossed. Each element can be either:

  • string: Will use the corresponding feature which must be of string type.
  • CategoricalColumn: Will use the transformed tensor produced by this column. Does not support hashed categorical column.
hash_bucket_size An int > 1. The number of buckets.
hash_key Specify the hash_key that will be used by the FingerprintCat64 function to combine the crosses fingerprints on SparseCrossOp (optional).

A CrossedColumn.

ValueError If len(keys) < 2.
ValueError If any of the keys is neither a string nor CategoricalColumn.
ValueError If any of the keys is HashedCategoricalColumn.
ValueError If hash_bucket_size < 1.