"""Gray encoding"""
from functools import partialmethod
import pandas as pd
from category_encoders import utils
from category_encoders.basen import BaseNEncoder
from typing import List
__author__ = 'paulwestenthanner'
[docs]class GrayEncoder(BaseNEncoder):
"""Gray encoding for categorical variables.
Gray encoding is a form of binary encoding where consecutive values only differ by a single bit.
Hence, gray encoding only makes sense for ordinal features.
This has benefits in privacy preserving data publishing.
Parameters
----------
verbose: int
integer indicating verbosity of the output. 0 for none.
cols: list
a list of columns to encode, if None, all string columns will be encoded.
drop_invariant: bool
boolean for whether or not to drop columns with 0 variance.
return_df: bool
boolean for whether to return a pandas DataFrame from transform (otherwise it will be a numpy array).
handle_unknown: str
options are 'error', 'return_nan', 'value', and 'indicator'. The default is 'value'. Warning: if indicator is used,
an extra column will be added in if the transform matrix has unknown categories. This can cause
unexpected changes in dimension in some cases.
handle_missing: str
options are 'error', 'return_nan', 'value', and 'indicator'. The default is 'value'. Warning: if indicator is used,
an extra column will be added in if the transform matrix has nan values. This can cause
unexpected changes in dimension in some cases.
Example
-------
>>> from category_encoders import GrayEncoder
>>> import pandas as pd
>>> from sklearn.datasets import fetch_openml
>>> bunch = fetch_openml(name="house_prices", as_frame=True)
>>> display_cols = ["Id", "MSSubClass", "MSZoning", "LotFrontage", "YearBuilt", "Heating", "CentralAir"]
>>> y = bunch.target
>>> X = pd.DataFrame(bunch.data, columns=bunch.feature_names)[display_cols]
>>> enc = GrayEncoder(cols=['CentralAir', 'Heating']).fit(X, y)
>>> numeric_dataset = enc.transform(X)
>>> print(numeric_dataset.info())
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 1460 entries, 0 to 1459
Data columns (total 10 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 Id 1460 non-null float64
1 MSSubClass 1460 non-null float64
2 MSZoning 1460 non-null object
3 LotFrontage 1201 non-null float64
4 YearBuilt 1460 non-null float64
5 Heating_0 1460 non-null int64
6 Heating_1 1460 non-null int64
7 Heating_2 1460 non-null int64
8 CentralAir_0 1460 non-null int64
9 CentralAir_1 1460 non-null int64
dtypes: float64(4), int64(5), object(1)
memory usage: 114.2+ KB
None
References
----------
.. [1] https://en.wikipedia.org/wiki/Gray_code
.. [2] Jun Zhang, Graham Cormode, Cecilia M. Procopiuc, Divesh Srivastava, and Xiaokui Xiao. 2017. PrivBayes:
Private Data Release via Bayesian Networks. ACM Trans. Database Syst. 42, 4, Article 25 (October 2017)
"""
encoding_relation = utils.EncodingRelation.ONE_TO_M
__init__ = partialmethod(BaseNEncoder.__init__, base=2)
@staticmethod
def gray_code(n, n_bit) -> List[int]:
gray = n ^ (n >> 1)
gray_formatted = "{0:0{1}b}".format(gray, n_bit)
return [int(bit) for bit in gray_formatted]
def _fit(self, X, y=None, **kwargs):
super(GrayEncoder, self)._fit(X, y, **kwargs)
gray_mapping = []
# convert binary mapping to Gray mapping and reorder
for col_to_encode in self.mapping:
col = col_to_encode["col"]
bin_mapping = col_to_encode["mapping"]
n_cols_out = bin_mapping.shape[1]
null_cond = (bin_mapping.index < 0) | (bin_mapping.isna().all(1))
map_null = bin_mapping[null_cond]
map_non_null = bin_mapping[~null_cond].copy()
ordinal_mapping = [m for m in self.ordinal_encoder.mapping if m.get("col") == col]
if len(ordinal_mapping) != 1:
raise ValueError("Cannot find ordinal encoder mapping of Gray encoder")
ordinal_mapping = ordinal_mapping[0]["mapping"]
reverse_ordinal_mapping = {v: k for k, v in ordinal_mapping.to_dict().items()}
map_non_null["orig_value"] = map_non_null.index.to_series().map(reverse_ordinal_mapping)
map_non_null = map_non_null.sort_values(by="orig_value")
gray_encoding = [self.gray_code(i + 1, n_cols_out) for i in range(map_non_null.shape[0])]
gray_encoding = pd.DataFrame(data=gray_encoding, index=map_non_null.index, columns=bin_mapping.columns)
gray_encoding = pd.concat([gray_encoding, map_null])
gray_mapping.append({"col": col, "mapping": gray_encoding})
self.mapping = gray_mapping