【category_encoders】分类特征编码方式

Table of Contents

import numpy as np 
import pandas as pd
import matplotlib.pyplot as plt 
import seaborn as sns
%matplotlib inline
plt.rcParams["font.sans-serif"] = ["FangSong"] 
plt.rcParams["axes.unicode_minus"] = False 
import warnings
warnings.filterwarnings("ignore")

import category_encoders as ce

非监督不分裂 ：‘OrdinalEncoder’,‘CountEncoder’
监督不分裂：‘TargetEncoder’,’ LeaveOneOutEncoder’,‘CatBoostEncoder’,‘WOEEncoder’
非监督分裂:‘OneHotEncoder’,‘Binary Encoder’

dir(ce)

['BackwardDifferenceEncoder',
 'BaseNEncoder',
 'BinaryEncoder',
 'CatBoostEncoder',
 'CountEncoder',
 'GLMMEncoder',
 'HashingEncoder',
 'HelmertEncoder',
 'JamesSteinEncoder',
 'LeaveOneOutEncoder',
 'MEstimateEncoder',
 'OneHotEncoder',
 'OrdinalEncoder',
 'PolynomialEncoder',
 'SumEncoder',
 'TargetEncoder',
 'WOEEncoder',
 '__all__',
 '__author__',
 '__builtins__',
 '__cached__',
 '__doc__',
 '__file__',
 '__loader__',
 '__name__',
 '__package__',
 '__path__',
 '__spec__',
 '__version__',
 'backward_difference',
 'basen',
 'binary',
 'cat_boost',
 'count',
 'glmm',
 'hashing',
 'helmert',
 'james_stein',
 'leave_one_out',
 'm_estimate',
 'one_hot',
 'ordinal',
 'polynomial',
 'sum_coding',
 'target_encoder',
 'utils',
 'woe']

X = pd.DataFrame(np.array([['male',10],['female', 20], ['male',10], 
                       ['female',20],['female',10],['female',30],['male',10]]),
             columns = ['Sex','Type'])
y = np.array([0,1,1,0,1,0,1])

X

	Sex	Type
0	male	10
1	female	20
2	male	10
3	female	20
4	female	10
5	female	30
6	male	10

OrdinalEncoder 序列编码

相当于sklearn中的LabelEncode
然而不是很好编码缺失值，应之前填充

encoder = ce.OrdinalEncoder(cols = ['Sex', 'Type']).fit(X,y)

encoder

OrdinalEncoder(cols=['Sex', 'Type'],
               mapping=[{'col': 'Sex', 'data_type': dtype('O'),
                         'mapping': male      1
female    2
NaN      -2
dtype: int64},
                        {'col': 'Type', 'data_type': dtype('O'),
                         'mapping': 10     1
20     2
30     3
NaN   -2
dtype: int64}])

encoder.transform(X) 

	Sex	Type
0	1	1
1	2	2
2	1	1
3	2	2
4	2	1
5	2	3
6	1	1

OneHotEncoder 独热编码

相当于pandas中get_dummies

encoder = ce.OneHotEncoder(cols = ['Sex', 'Type'],drop_invariant= True).fit(X,y)

encoder.transform(X) 

	Sex_1	Sex_2	Type_1	Type_2	Type_3
0	1	0	1	0	0
1	0	1	0	1	0
2	1	0	1	0	0
3	0	1	0	1	0
4	0	1	1	0	0
5	0	1	0	0	1
6	1	0	1	0	0

TargetEncoder 目标编码

link：https://zhuanlan.zhihu.com/p/119093636
[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-WOwbyUfY-1603970994566)(attachment:image.png)]

encoder = ce.TargetEncoder(cols = ['Sex', 'Type'],drop_invariant= True).fit(X,y)

encoder.transform(X)

	Sex	Type
0	0.655314	0.741531
1	0.503388	0.519210
2	0.655314	0.741531
3	0.503388	0.519210
4	0.503388	0.741531
5	0.503388	0.571429
6	0.655314	0.741531

Binary Encoder二进制编码

encoder = ce.BinaryEncoder(cols = ['Sex', 'Type']).fit(X,y)

encoder.transform(X)

	Sex_0	Sex_1	Type_0	Type_1	Type_2
0	0	1	0	0	1
1	1	0	0	1	0
2	0	1	0	0	1
3	1	0	0	1	0
4	1	0	0	0	1
5	1	0	0	1	1
6	0	1	0	0	1

BaseNEncoder 贝叶斯编码

encoder = ce.BaseNEncoder(cols = ['Sex', 'Type']).fit(X,y)

encoder.transform(X)

	Sex_0	Sex_1	Type_0	Type_1	Type_2
0	0	1	0	0	1
1	1	0	0	1	0
2	0	1	0	0	1
3	1	0	0	1	0
4	1	0	0	0	1
5	1	0	0	1	1
6	0	1	0	0	1

LeaveOneOutEncoder 留一法

类似目标编码 

encoder = ce.LeaveOneOutEncoder(cols = ['Sex', 'Type']).fit(X,y)

encoder.transform(X)

	Sex	Type
0	0.666667	0.750000
1	0.500000	0.500000
2	0.666667	0.750000
3	0.500000	0.500000
4	0.500000	0.750000
5	0.500000	0.571429
6	0.666667	0.750000

HashingEncoder 哈希编码

encoder = ce.HashingEncoder(cols = ['Sex', 'Type']).fit(X,y)
encoder.transform(X)

	col_0	col_1	col_2	col_3	col_4	col_5	col_6	col_7
0	1	0	0	0	0	1	0	0
1	0	0	0	0	1	1	0	0
2	1	0	0	0	0	1	0	0
3	0	0	0	0	1	1	0	0
4	1	0	0	0	0	1	0	0
5	0	0	0	0	0	1	0	1
6	1	0	0	0	0	1	0	0

CatBoostEncoder catboost目标编码

encoder = ce.CatBoostEncoder(cols = ['Sex', 'Type']).fit(X,y)
encoder.transform(X)

	Sex	Type
0	0.642857	0.714286
1	0.514286	0.523810
2	0.642857	0.714286
3	0.514286	0.523810
4	0.514286	0.714286
5	0.514286	0.571429
6	0.642857	0.714286

CountEncoder 频率编码

encoder = ce.CountEncoder(cols = ['Sex', 'Type']).fit(X,y)
encoder.transform(X)

	Sex	Type
0	3	4
1	4	2
2	3	4
3	4	2
4	4	4
5	4	1
6	3	4

WOEEncoder 证据权重编码

encoder = ce.WOEEncoder(cols = ['Sex', 'Type']).fit(X,y)
encoder.transform(X)

	Sex	Type
0	0.223144	0.510826
1	-0.182322	-0.182322
2	0.223144	0.510826
3	-0.182322	-0.182322
4	-0.182322	0.510826
5	-0.182322	0.000000
6	0.223144	0.510826

dir(ce)

['BackwardDifferenceEncoder',
 'BaseNEncoder',
 'BinaryEncoder',
 'CatBoostEncoder',
 'CountEncoder',
 'GLMMEncoder',
 'HashingEncoder',
 'HelmertEncoder',
 'JamesSteinEncoder',
 'LeaveOneOutEncoder',
 'MEstimateEncoder',
 'OneHotEncoder',
 'OrdinalEncoder',
 'PolynomialEncoder',
 'SumEncoder',
 'TargetEncoder',
 'WOEEncoder',
 '__all__',
 '__author__',
 '__builtins__',
 '__cached__',
 '__doc__',
 '__file__',
 '__loader__',
 '__name__',
 '__package__',
 '__path__',
 '__spec__',
 '__version__',
 'backward_difference',
 'basen',
 'binary',
 'cat_boost',
 'count',
 'glmm',
 'hashing',
 'helmert',
 'james_stein',
 'leave_one_out',
 'm_estimate',
 'one_hot',
 'ordinal',
 'polynomial',
 'sum_coding',
 'target_encoder',
 'utils',
 'woe']