View Source Scholar.Preprocessing.OrdinalEncoder (Scholar v0.4.0)

Implements encoder that converts integer value (substitute of categorical data in tensors) into other integer value. The values assigned starts from 0 and go up to num_categories - 1. They are maintained in sorted manner. This means that for x < y => encoded_value(x) < encoded_value(y).

Currently the module supports only 1D tensors.

Summary

Functions

fit(tensor, opts \\ [])

Fit the ordinal encoder to provided data. The labels are assigned in a sorted manner.

fit_transform(tensor)

Apply encoding on the provided tensor directly. It's equivalent to fit/2 and then transform/2 on the same data.

inverse_transform(ordinal_encoder, encoded_tensor)

Decodes tensor elements into original categories seen during fitting.

transform(arg1, tensor)

Encodes tensor elements into integers from range 0 to :num_categories - 1 or -1 if the value did not occur in fitting process.

Functions

fit(tensor, opts \\ [])

Fit the ordinal encoder to provided data. The labels are assigned in a sorted manner.

Options

:num_categories (pos_integer/0) - Required. The number of categories to be encoded.

Examples

iex> tensor = Nx.tensor([3, 2, 4, 56, 2, 4, 2])
iex> Scholar.Preprocessing.OrdinalEncoder.fit(tensor, num_categories: 4)
%Scholar.Preprocessing.OrdinalEncoder{
  categories: Nx.tensor([2, 3, 4, 56])
}

fit_transform(tensor)

Apply encoding on the provided tensor directly. It's equivalent to fit/2 and then transform/2 on the same data.

Examples

iex> tensor = Nx.tensor([3, 2, 4, 56, 2, 4, 2])
iex> Scholar.Preprocessing.OrdinalEncoder.fit_transform(tensor)
#Nx.Tensor<
  u64[7]
  [1, 0, 2, 3, 0, 2, 0]
>

inverse_transform(ordinal_encoder, encoded_tensor)

Decodes tensor elements into original categories seen during fitting.

Examples

iex> tensor = Nx.tensor([3, 2, 4, 56, 2, 4, 2]) iex> encoder = Scholar.Preprocessing.OrdinalEncoder.fit(tensor, num_categories: 4) iex> encoded = Nx.tensor([1, 0, 2, 3, 1, 0, 2]) iex> Scholar.Preprocessing.OrdinalEncoder.inverse_transform(encoder, encoded) Nx.tensor([3, 2, 4, 56, 3, 2, 4])

transform(arg1, tensor)

Encodes tensor elements into integers from range 0 to :num_categories - 1 or -1 if the value did not occur in fitting process.

Examples

iex> tensor = Nx.tensor([3, 2, 4, 56, 2, 4, 2])
iex> encoder = Scholar.Preprocessing.OrdinalEncoder.fit(tensor, num_categories: 4)
iex> Scholar.Preprocessing.OrdinalEncoder.transform(encoder, tensor)
#Nx.Tensor<
  s32[7]
  [1, 0, 2, 3, 0, 2, 0]
>

iex> tensor = Nx.tensor([3, 2, 4, 56, 2, 4, 2])
iex> encoder = Scholar.Preprocessing.OrdinalEncoder.fit(tensor, num_categories: 4)
iex> new_tensor = Nx.tensor([2, 3, 4, 5, 4, 56, 2])
iex> Scholar.Preprocessing.OrdinalEncoder.transform(encoder, new_tensor)
#Nx.Tensor<
  s32[7]
  [0, 1, 2, -1, 2, 3, 0]
>