SMOTE抽样 数据不平衡的问题

from imblearn.over_sampling import SMOTE

import pandas as pd 

C:\ProgramData\Anaconda3\lib\importlib\_bootstrap.py:219: RuntimeWarning: numpy.ufunc size changed, may indicate binary incompatibility. Expected 192 from C header, got 216 from PyObject
  return f(*args, **kwds)
C:\ProgramData\Anaconda3\lib\importlib\_bootstrap.py:219: RuntimeWarning: numpy.ufunc size changed, may indicate binary incompatibility. Expected 192 from C header, got 216 from PyObject
  return f(*args, **kwds)

df = pd.read_csv('base_done.csv')

data = df[:20].iloc[:,1:10]
data

	sex	age	provider	level	verified	using_time	regist_type	card_a_cnt	card_b_cnt
0	0	24853	0	1	0	24713	1	24719	24712
1	1	25011	0	1	0	24743	7	24712	24712
2	1	24877	0	2	0	24744	7	24719	24725
3	0	24925	0	2	0	24715	1	24712	24712
4	1	24877	2	1	0	24706	3	24712	24712
5	0	24944	0	2	0	24720	1	24719	24712
6	0	24840	0	2	0	24727	1	24712	24712
7	0	24944	0	2	0	24709	1	24712	24712
8	0	24908	0	2	0	24730	1	24712	24712
9	1	24956	0	2	0	24741	7	24719	24719
10	0	24920	0	2	0	24741	7	24719	24719
11	0	24871	0	2	0	24722	2	24712	24712
12	1	24889	0	2	0	24725	1	24719	24719
13	1	24865	2	2	0	24710	3	24712	24712
14	0	24944	0	2	0	24727	1	24712	24712
15	1	24931	0	2	0	24733	7	24712	24719
16	0	24963	0	1	0	24721	2	24712	24712
17	0	24877	0	2	0	24727	1	24712	24712
18	0	24901	0	2	0	24733	7	24725	24719
19	0	24859	2	1	0	24713	3	24712	24712

X = data.drop(columns='provider').values

y = data.provider

data.provider.value_counts()

0    17
2     3
Name: provider, dtype: int64

sm = SMOTE(sampling_strategy={0:17,2:15},k_neighbors=2)  # sampling_strategy 抽样策略，默认为1:1， k = knn的
X_res, y_res = sm.fit_resample(X, y)  # 

y_res.value_counts()

0    17
2    15
Name: provider, dtype: int64

X_res

	sex	age	level	verified	using_time	regist_type	card_a_cnt	card_b_cnt
0	0	24853	1	0	24713	1	24719	24712
1	1	25011	1	0	24743	7	24712	24712
2	1	24877	2	0	24744	7	24719	24725
3	0	24925	2	0	24715	1	24712	24712
4	1	24877	1	0	24706	3	24712	24712
5	0	24944	2	0	24720	1	24719	24712
6	0	24840	2	0	24727	1	24712	24712
7	0	24944	2	0	24709	1	24712	24712
8	0	24908	2	0	24730	1	24712	24712
9	1	24956	2	0	24741	7	24719	24719
10	0	24920	2	0	24741	7	24719	24719
11	0	24871	2	0	24722	2	24712	24712
12	1	24889	2	0	24725	1	24719	24719
13	1	24865	2	0	24710	3	24712	24712
14	0	24944	2	0	24727	1	24712	24712
15	1	24931	2	0	24733	7	24712	24719
16	0	24963	1	0	24721	2	24712	24712
17	0	24877	2	0	24727	1	24712	24712
18	0	24901	2	0	24733	7	24725	24719
19	0	24859	1	0	24713	3	24712	24712
20	0	24860	1	0	24712	3	24712	24712
21	0	24871	1	0	24708	3	24712	24712
22	0	24872	1	0	24707	3	24712	24712
23	1	24865	1	0	24709	3	24712	24712
24	0	24863	1	0	24710	3	24712	24712
25	0	24859	1	0	24712	3	24712	24712
26	1	24870	1	0	24708	3	24712	24712
27	1	24875	1	0	24706	3	24712	24712
28	1	24872	1	0	24707	3	24712	24712
29	0	24873	1	0	24707	3	24712	24712
30	0	24862	1	0	24711	3	24712	24712
31	0	24859	1	0	24712	3	24712	24712
32	0	24873	1	0	24707	3	24712	24712
33	1	24866	1	0	24709	3	24712	24712