处理分类变量

Python 中的机器学习特征工程

Robert O'Callaghan

Director of Data Science, Ordergroove

编码分类特征

One-hot 编码
哑变量编码

One-hot 编码

pd.get_dummies(df, columns=['Country'], 
               prefix='C')

    C_France    C_India    C_UK    C_USA
0          0          1       0        0
1          0          0       0        1
2          0          0       1        0
3          0          0       1        0
4          1          0       0        0

哑变量编码

pd.get_dummies(df, columns=['Country'],
               drop_first=True, prefix='C')

     C_India    C_UK    C_USA
0          1       0        0
1          0       0        1
2          0       1        0
3          0       1        0
4          0       0        0

One-hot 与哑变量

One-hot 编码：特征易于解释
哑变量编码：去重后保留必要信息

Index	Sex
0	Male
1	Female
2	Male

Index	Male	Female
0	1	0
1	0	1
2	1	0

Index	Male
0	1
1	0
2	1

限制列数量

counts = df['Country'].value_counts()
print(counts)

'USA'      8
'UK'       6
'India'    2
'France'   1
Name: Country, dtype: object

限制列数量

mask = df['Country'].isin(counts[counts < 5].index)

df['Country'][mask] = 'Other'

print(pd.value_counts(colors))

'USA'      8
'UK'       6
'Other'    3
Name: Country, dtype: object

现在您会处理分类变量了

Python 中的机器学习特征工程