一、项目介绍

背景

以金融风控中的个人信贷为背景，根据贷款申请人的数据信息预测其是否有违约的可能，以此判断是否通过此项贷款，这是一个典型的分类问题。

具体的列名含义

            id 为贷款清单分配的唯一信用证标识loanAmnt 贷款金额term 贷款期限（year）interestRate 贷款利率installment 分期付款金额grade 贷款等级subGrade 贷款等级之子级employmentTitle 就业职称employmentLength 就业年限（年）homeOwnership 借款人在登记时提供的房屋所有权状况annualIncome 年收入verificationStatus 验证状态issueDate 贷款发放的月份purpose 借款人在贷款申请时的贷款用途类别postCode 借款人在贷款申请中提供的邮政编码的前3位数字regionCode 地区编码dti 债务收入比delinquency_2years 借款人过去2年信用档案中逾期30天以上的违约事件数ficoRangeLow 借款人在贷款发放时的fico所属的下限范围ficoRangeHigh 借款人在贷款发放时的fico所属的上限范围openAcc 借款人信用档案中未结信用额度的数量pubRec 贬损公共记录的数量pubRecBankruptcies 公开记录清除的数量revolBal 信贷周转余额合计revolUtil 循环额度利用率，或借款人使用的相对于所有可用循环信贷的信贷金额totalAcc 借款人信用档案中当前的信用额度总数initialListStatus 贷款的初始列表状态applicationType 表明贷款是个人申请还是与两个共同借款人的联合申请earliesCreditLine 借款人最早报告的信用额度开立的月份title 借款人提供的贷款名称policyCode 公开可用的策略代码=1新产品不公开可用的策略代码=2n系列匿名特征 匿名特征n0-n14，为一些贷款人行为计数特征的处理

二、数据准备

导入相关库

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
import datetime
from sklearn.model_selection import cross_val_score,train_test_split,GridSearchCV
from sklearn.linear_model import LogisticRegression
from sklearn.ensemble import RandomForestClassifier
import warnings
warnings.filterwarnings('ignore')##### 取消pandas最大列显示限制
pd.options.display.max_columns = None

获取数据

train = pd.read_csv('../data/贷款违约预测/train.csv')

三、数据分析

3.1 总体了解数据

train.shape

(800000, 47)

train.columns

Index(['id', 'loanAmnt', 'term', 'interestRate', 'installment', 'grade','subGrade', 'employmentTitle', 'employmentLength', 'homeOwnership','annualIncome', 'verificationStatus', 'issueDate', 'isDefault','purpose', 'postCode', 'regionCode', 'dti', 'delinquency_2years','ficoRangeLow', 'ficoRangeHigh', 'openAcc', 'pubRec','pubRecBankruptcies', 'revolBal', 'revolUtil', 'totalAcc','initialListStatus', 'applicationType', 'earliesCreditLine', 'title','policyCode', 'n0', 'n1', 'n2', 'n3', 'n4', 'n5', 'n6', 'n7', 'n8','n9', 'n10', 'n11', 'n12', 'n13', 'n14'],dtype='object')

train.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 800000 entries, 0 to 799999
Data columns (total 47 columns):#   Column              Non-Null Count   Dtype
---  ------              --------------   -----  0   id                  800000 non-null  int64  1   loanAmnt            800000 non-null  float642   term                800000 non-null  int64  3   interestRate        800000 non-null  float644   installment         800000 non-null  float645   grade               800000 non-null  object 6   subGrade            800000 non-null  object 7   employmentTitle     799999 non-null  float648   employmentLength    753201 non-null  object 9   homeOwnership       800000 non-null  int64  10  annualIncome        800000 non-null  float6411  verificationStatus  800000 non-null  int64  12  issueDate           800000 non-null  object 13  isDefault           800000 non-null  int64  14  purpose             800000 non-null  int64  15  postCode            799999 non-null  float6416  regionCode          800000 non-null  int64  17  dti                 799761 non-null  float6418  delinquency_2years  800000 non-null  float6419  ficoRangeLow        800000 non-null  float6420  ficoRangeHigh       800000 non-null  float6421  openAcc             800000 non-null  float6422  pubRec              800000 non-null  float6423  pubRecBankruptcies  799595 non-null  float6424  revolBal            800000 non-null  float6425  revolUtil           799469 non-null  float6426  totalAcc            800000 non-null  float6427  initialListStatus   800000 non-null  int64  28  applicationType     800000 non-null  int64  29  earliesCreditLine   800000 non-null  object 30  title               799999 non-null  float6431  policyCode          800000 non-null  float6432  n0                  759730 non-null  float6433  n1                  759730 non-null  float6434  n2                  759730 non-null  float6435  n3                  759730 non-null  float6436  n4                  766761 non-null  float6437  n5                  759730 non-null  float6438  n6                  759730 non-null  float6439  n7                  759730 non-null  float6440  n8                  759729 non-null  float6441  n9                  759730 non-null  float6442  n10                 766761 non-null  float6443  n11                 730248 non-null  float6444  n12                 759730 non-null  float6445  n13                 759730 non-null  float6446  n14                 759730 non-null  float64
dtypes: float64(33), int64(9), object(5)
memory usage: 286.9+ MB

train.describe()

	id	loanAmnt	term	interestRate	installment	employmentTitle	homeOwnership	annualIncome	verificationStatus	isDefault	purpose	postCode	regionCode	dti	delinquency_2years	ficoRangeLow	ficoRangeHigh	openAcc	pubRec	pubRecBankruptcies	revolBal	revolUtil	totalAcc	initialListStatus	applicationType	title	policyCode	n0	n1	n2	n3	n4	n5	n6	n7	n8	n9	n10	n11	n12	n13	n14
count	800000.000000	800000.000000	800000.000000	800000.000000	800000.000000	799999.000000	800000.000000	8.000000e+05	800000.000000	800000.000000	800000.000000	799999.000000	800000.000000	799761.000000	800000.000000	800000.000000	800000.000000	800000.000000	800000.000000	799595.000000	8.000000e+05	799469.000000	800000.000000	800000.000000	800000.000000	799999.000000	800000.0	759730.000000	759730.000000	759730.000000	759730.000000	766761.000000	759730.000000	759730.000000	759730.000000	759729.000000	759730.000000	766761.000000	730248.000000	759730.000000	759730.000000	759730.000000
mean	399999.500000	14416.818875	3.482745	13.238391	437.947723	72005.351714	0.614213	7.613391e+04	1.009683	0.199513	1.745982	258.535648	16.385758	18.284557	0.318239	696.204081	700.204226	11.598020	0.214915	0.134163	1.622871e+04	51.790734	24.998861	0.416953	0.019267	1754.113589	1.0	0.511932	3.642330	5.642648	5.642648	4.735641	8.107937	8.575994	8.282953	14.622488	5.592345	11.643896	0.000815	0.003384	0.089366	2.178606
std	230940.252015	8716.086178	0.855832	4.765757	261.460393	106585.640204	0.675749	6.894751e+04	0.782716	0.399634	2.367453	200.037446	11.036679	11.150155	0.880325	31.865995	31.866674	5.475286	0.606467	0.377471	2.245802e+04	24.516126	11.999201	0.493055	0.137464	7941.474040	0.0	1.333266	2.246825	3.302810	3.302810	2.949969	4.799210	7.400536	4.561689	8.124610	3.216184	5.484104	0.030075	0.062041	0.509069	1.844377
min	0.000000	500.000000	3.000000	5.310000	15.690000	0.000000	0.000000	0.000000e+00	0.000000	0.000000	0.000000	0.000000	0.000000	-1.000000	0.000000	630.000000	634.000000	0.000000	0.000000	0.000000	0.000000e+00	0.000000	2.000000	0.000000	0.000000	0.000000	1.0	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	1.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000
25%	199999.750000	8000.000000	3.000000	9.750000	248.450000	427.000000	0.000000	4.560000e+04	0.000000	0.000000	0.000000	103.000000	8.000000	11.790000	0.000000	670.000000	674.000000	8.000000	0.000000	0.000000	5.944000e+03	33.400000	16.000000	0.000000	0.000000	0.000000	1.0	0.000000	2.000000	3.000000	3.000000	3.000000	5.000000	4.000000	5.000000	9.000000	3.000000	8.000000	0.000000	0.000000	0.000000	1.000000
50%	399999.500000	12000.000000	3.000000	12.740000	375.135000	7755.000000	1.000000	6.500000e+04	1.000000	0.000000	0.000000	203.000000	14.000000	17.610000	0.000000	690.000000	694.000000	11.000000	0.000000	0.000000	1.113200e+04	52.100000	23.000000	0.000000	0.000000	1.000000	1.0	0.000000	3.000000	5.000000	5.000000	4.000000	7.000000	7.000000	7.000000	13.000000	5.000000	11.000000	0.000000	0.000000	0.000000	2.000000
75%	599999.250000	20000.000000	3.000000	15.990000	580.710000	117663.500000	1.000000	9.000000e+04	2.000000	0.000000	4.000000	395.000000	22.000000	24.060000	0.000000	710.000000	714.000000	14.000000	0.000000	0.000000	1.973400e+04	70.700000	32.000000	1.000000	0.000000	5.000000	1.0	0.000000	5.000000	7.000000	7.000000	6.000000	11.000000	11.000000	10.000000	19.000000	7.000000	14.000000	0.000000	0.000000	0.000000	3.000000
max	799999.000000	40000.000000	5.000000	30.990000	1715.420000	378351.000000	5.000000	1.099920e+07	2.000000	1.000000	13.000000	940.000000	50.000000	999.000000	39.000000	845.000000	850.000000	86.000000	86.000000	12.000000	2.904836e+06	892.300000	162.000000	1.000000	1.000000	61680.000000	1.0	51.000000	33.000000	63.000000	63.000000	49.000000	70.000000	132.000000	79.000000	128.000000	45.000000	82.000000	4.000000	4.000000	39.000000	30.000000

# 查看数据集中特征缺失值的特征数
train.isnull().any().sum()

# 具体的查看缺失特征数量并可视化
missing = train.isnull().sum()
missing = missing[missing > 0]
missing.sort_values(inplace = True)
missing.plot.bar();

[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-lUzXSsK9-1615121853776)(output_13_0.png)]

# 查看训练集测试集中特征属性只有一值的特征
fea = [col for col in train.columns if train[col].nunique() <=1]
fea

['policyCode']

# 查看特征的数值类型有哪些，对象类型有哪些
numerical_fea = list(train.select_dtypes(exclude=['object']).columns)
category_fea = list(filter(lambda x:x not in numerical_fea,list(train.columns)))
print('数值类型特征有{}个，分别为{}：'.format(len(numerical_fea),numerical_fea))
print()
print('对象类型特征有{}个，分别为{}：'.format(len(category_fea),category_fea))

数值类型特征有42个，分别为['id', 'loanAmnt', 'term', 'interestRate', 'installment', 'employmentTitle', 'homeOwnership', 'annualIncome', 'verificationStatus', 'isDefault', 'purpose', 'postCode', 'regionCode', 'dti', 'delinquency_2years', 'ficoRangeLow', 'ficoRangeHigh', 'openAcc', 'pubRec', 'pubRecBankruptcies', 'revolBal', 'revolUtil', 'totalAcc', 'initialListStatus', 'applicationType', 'title', 'policyCode', 'n0', 'n1', 'n2', 'n3', 'n4', 'n5', 'n6', 'n7', 'n8', 'n9', 'n10', 'n11', 'n12', 'n13', 'n14']：对象类型特征有5个，分别为['grade', 'subGrade', 'employmentLength', 'issueDate', 'earliesCreditLine']：

# 划分数值型变量中的连续变量和离散型变量
numerical_noserial_fea = []
numerical_serial_fea = []for fea in numerical_fea:temp = train[fea].nunique()if temp <= 10:numerical_noserial_fea.append(fea)continuenumerical_serial_fea.append(fea)print('数值连续型变量特征有：',numerical_serial_fea)
print()
print('数值离散型变量特征有：',numerical_noserial_fea)

数值连续型变量特征有： ['id', 'loanAmnt', 'interestRate', 'installment', 'employmentTitle', 'annualIncome', 'purpose', 'postCode', 'regionCode', 'dti', 'delinquency_2years', 'ficoRangeLow', 'ficoRangeHigh', 'openAcc', 'pubRec', 'pubRecBankruptcies', 'revolBal', 'revolUtil', 'totalAcc', 'title', 'n0', 'n1', 'n2', 'n3', 'n4', 'n5', 'n6', 'n7', 'n8', 'n9', 'n10', 'n13', 'n14']数值离散型变量特征有： ['term', 'homeOwnership', 'verificationStatus', 'isDefault', 'initialListStatus', 'applicationType', 'policyCode', 'n11', 'n12']

3.2 数值离散型变量分析

for fea in numerical_noserial_fea:print('离散型变量：',fea)print(train[fea].value_counts())print()print()

离散型变量： term
3    606902
5    193098
Name: term, dtype: int64离散型变量： homeOwnership
0    395732
1    317660
2     86309
3       185
5        81
4        33
Name: homeOwnership, dtype: int64离散型变量： verificationStatus
1    309810
2    248968
0    241222
Name: verificationStatus, dtype: int64离散型变量： isDefault
0    640390
1    159610
Name: isDefault, dtype: int64离散型变量： initialListStatus
0    466438
1    333562
Name: initialListStatus, dtype: int64离散型变量： applicationType
0    784586
1     15414
Name: applicationType, dtype: int64离散型变量： policyCode
1.0    800000
Name: policyCode, dtype: int64离散型变量： n11
0.0    729682
1.0       540
2.0        24
4.0         1
3.0         1
Name: n11, dtype: int64离散型变量： n12
0.0    757315
1.0      2281
2.0       115
3.0        16
4.0         3
Name: n12, dtype: int64

3.3 数值连续型变量分析

f = pd.melt(train, value_vars=numerical_serial_fea)
g = sns.FacetGrid(f, col="variable",  col_wrap=4, sharex=False, sharey=False)
g = g.map(sns.distplot, "value")

[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-GxlERoUd-1615121853777)(output_20_0.png)]

3.4 非数值类别型变量分析

for fea in category_fea:print('非数值类别型变量：',fea)print(train[fea].value_counts())print()

非数值类别型变量： grade
B    233690
C    227118
A    139661
D    119453
E     55661
F     19053
G      5364
Name: grade, dtype: int64非数值类别型变量： subGrade
C1    50763
B4    49516
B5    48965
B3    48600
C2    47068
C3    44751
C4    44272
B2    44227
B1    42382
C5    40264
A5    38045
A4    30928
D1    30538
D2    26528
A1    25909
D3    23410
A3    22655
A2    22124
D4    21139
D5    17838
E1    14064
E2    12746
E3    10925
E4     9273
E5     8653
F1     5925
F2     4340
F3     3577
F4     2859
F5     2352
G1     1759
G2     1231
G3      978
G4      751
G5      645
Name: subGrade, dtype: int64非数值类别型变量： employmentLength
10+ years    262753
2 years       72358
< 1 year      64237
3 years       64152
1 year        52489
5 years       50102
4 years       47985
6 years       37254
8 years       36192
7 years       35407
9 years       30272
Name: employmentLength, dtype: int64非数值类别型变量： issueDate
2016-03-01    29066
2015-10-01    25525
2015-07-01    24496
2015-12-01    23245
2014-10-01    21461...
2007-08-01       23
2007-07-01       21
2008-09-01       19
2007-09-01        7
2007-06-01        1
Name: issueDate, Length: 139, dtype: int64非数值类别型变量： earliesCreditLine
Aug-2001    5567
Sep-2003    5403
Aug-2002    5403
Oct-2001    5258
Aug-2000    5246...
Oct-1954       1
Jan-1944       1
May-1957       1
Nov-1954       1
Nov-1953       1
Name: earliesCreditLine, Length: 720, dtype: int64

三、特征工程

3.1 特征预处理

3.1.1缺失值填充

# 按照平均数填充连续型数值型特征
train[numerical_fea] = train[numerical_fea].fillna(train[numerical_fea].median())
# 按照众数填充类别型特征
train[category_fea].fillna(train[category_fea].mode(),inplace=True)
train['employmentLength'].fillna('10+ years',inplace=True)
train.isnull().any().sum()

3.1.2 对象型类别特征进行预处理

# 时间格式处理
train['issueDate'] = pd.to_datetime(train['issueDate'],format='%Y-%m-%d')
startdate = datetime.datetime.strptime('2007-06-01','%Y-%m-%d')
train['issueDate'] = train['issueDate'].apply(lambda x: x-startdate).dt.days
train['issueDate'].value_counts()

3196    29066
3044    25525
2952    24496
3105    23245
2679    21461...
61         23
30         21
458        19
92          7
0           1
Name: issueDate, Length: 139, dtype: int64

# employmentLength预处理
def employmentLength_to_int(s):if pd.isnull(s):return selse:return np.int8(s.split()[0])train['employmentLength'].replace(to_replace='10+ years', value='10 years', inplace=True)
train['employmentLength'].replace('< 1 year', '0 years', inplace=True)
train['employmentLength'] = train['employmentLength'].apply(employmentLength_to_int)
train['employmentLength'].value_counts()

10    309552
2      72358
0      64237
3      64152
1      52489
5      50102
4      47985
6      37254
8      36192
7      35407
9      30272
Name: employmentLength, dtype: int64

# 对earliesCreditLine进行预处理
train['earliesCreditLine'] = train['earliesCreditLine'].apply(lambda x:int(x[-4:]))
train['earliesCreditLine'].value_counts()

2001    53194
2002    51060
2003    50649
2000    50624
2004    49280...
1954        5
1953        5
1950        5
1946        2
1944        1
Name: earliesCreditLine, Length: 68, dtype: int64

# grade预处理
train['grade'] = train['grade'].map({'A':1,'B':2,'C':3,'D':4,'E':5,'F':6,'G':7})
train['grade'].value_counts()

2    233690
3    227118
1    139661
4    119453
5     55661
6     19053
7      5364
Name: grade, dtype: int64

# subGrade预处理
train['subGrade'] = train['subGrade'].map({'A1':1,'A2':2,'A3':3,'A4':4,'A5':5,'B1':6,'B2':7,'B3':8,'B4':9,'B5':10,'C1':11,'C2':12,'C3':13,'C4':14,'C5':15
,'D1':16,'D2':17,'D3':18,'D4':19,'D5':20,'E1':21,'E2':22,'E3':23,'E4':24,'E5':25,'F1':26,'F2':27,'F3':28,'F4':29,'F5':30
,'G1':31,'G2':32,'G3':33,'G4':34,'G5':35})
train['subGrade'].value_counts()

11    50763
9     49516
10    48965
8     48600
12    47068
13    44751
14    44272
7     44227
6     42382
15    40264
5     38045
4     30928
16    30538
17    26528
1     25909
18    23410
3     22655
2     22124
19    21139
20    17838
21    14064
22    12746
23    10925
24     9273
25     8653
26     5925
27     4340
28     3577
29     2859
30     2352
31     1759
32     1231
33      978
34      751
35      645
Name: subGrade, dtype: int64

3.1.3 数值离散特征处理

# 独热编码
temp = ['subGrade','homeOwnership','verificationStatus','purpose','regionCode']
data = pd.get_dummies(train,columns=temp,drop_first=True)
data.head()

	id	loanAmnt	term	interestRate	installment	grade	employmentTitle	employmentLength	annualIncome	issueDate	isDefault	postCode	dti	ficoRangeLow	ficoRangeHigh	openAcc	revolBal	revolUtil	totalAcc	initialListStatus	earliesCreditLine	title	policyCode	n0	n1	n2	n3	n4	n5	n6	n7	n8	n9	n10	n14	subGrade_4	subGrade_12	subGrade_17	subGrade_18	subGrade_22	homeOwnership_1	homeOwnership_2	verificationStatus_1	verificationStatus_2	purpose_1	purpose_4	purpose_10	regionCode_11	regionCode_14	regionCode_18	regionCode_21	regionCode_32
0	0	35000.0	5	19.52	917.97	5	320.0	2	110000.0	2587	1	137.0	17.05	730.0	734.0	7.0	24178.0	48.9	27.0	0	2001	1.0	1.0	0.0	2.0	2.0	2.0	4.0	9.0	8.0	4.0	12.0	2.0	7.0	2.0	0	0	0	0	1	0	1	0	1	1	0	0	0	0	0	0	1
1	1	18000.0	5	18.49	461.90	4	219843.0	5	46000.0	1888	0	156.0	27.83	700.0	704.0	13.0	15096.0	38.9	18.0	1	2002	1723.0	1.0	0.0	3.0	5.0	5.0	10.0	7.0	7.0	7.0	13.0	5.0	13.0	2.0	0	0	1	0	0	0	0	0	1	0	0	0	0	0	1	0	0
2	2	12000.0	5	16.99	298.17	4	31698.0	8	74000.0	3044	0	337.0	22.77	675.0	679.0	11.0	4606.0	51.8	27.0	0	2006	0.0	1.0	0.0	0.0	3.0	3.0	0.0	0.0	21.0	4.0	5.0	3.0	11.0	4.0	0	0	0	1	0	0	0	0	1	0	0	0	0	1	0	0	0
3	3	11000.0	3	7.26	340.96	1	46854.0	10	118000.0	2983	0	148.0	17.21	685.0	689.0	9.0	9948.0	52.6	28.0	1	1999	4.0	1.0	6.0	4.0	6.0	6.0	4.0	16.0	4.0	7.0	21.0	6.0	9.0	1.0	1	0	0	0	0	1	0	1	0	0	1	0	1	0	0	0	0
4	4	3000.0	3	12.99	101.07	3	54.0	10	29000.0	3196	0	301.0	32.16	690.0	694.0	12.0	2942.0	32.0	27.0	0	1977	11.0	1.0	1.0	2.0	7.0	7.0	2.0	4.0	9.0	10.0	15.0	7.0	12.0	4.0	0	1	0	0	0	1	0	0	1	0	0	1	0	0	0	1	0

3.1.4 连续型特征处理

# 异常值处理
def find_outliers_by_3segama(fea):data_std = np.std(data[fea])data_mean = np.mean(data[fea])outliers_cut_off = data_std * 3lower_rule = data_mean - outliers_cut_offupper_rule = data_mean + outliers_cut_offdata[fea] = data[fea].apply(lambda x:np.nan if x > upper_rule or x < lower_rule else x)return datafor fea in data.columns:if data[fea].nunique() > 10:find_outliers_by_3segama(fea)continuedata.dropna(axis=0,how='any',inplace=True)
data.shape

(626125, 146)

# 数据分桶
for fea in data.columns:if data[fea].nunique() > 10:if data[fea].max()-data[fea].min() >100000:data[fea] = np.floor_divide(data[fea], 10000)elif data[fea].max()-data[fea].min() >10000:data[fea] = np.floor_divide(data[fea], 1000)elif data[fea].max()-data[fea].min() >1000:data[fea] = np.floor_divide(data[fea], 100)elif data[fea].max()-train[fea].min() >100:data[fea] = np.floor_divide(data[fea], 10)
data.head()

	loanAmnt	term	interestRate	installment	grade	employmentTitle	employmentLength	annualIncome	issueDate	isDefault	postCode	dti	ficoRangeLow	ficoRangeHigh	openAcc	revolBal	revolUtil	totalAcc	initialListStatus	earliesCreditLine	title	policyCode	n0	n1	n2	n3	n4	n5	n6	n7	n8	n9	n10	n14	subGrade_4	subGrade_12	subGrade_17	subGrade_18	subGrade_22	homeOwnership_1	homeOwnership_2	verificationStatus_2	purpose_1	purpose_10	regionCode_14	regionCode_18	regionCode_21	regionCode_32
0	35.0	5	19.52	9.0	5	0.0	2	11.0	25.0	1	13.0	17.05	73.0	73.0	7.0	24.0	4.0	27.0	0	2001.0	0.0	1.0	0.0	2.0	2.0	2.0	4.0	9.0	8.0	4.0	12.0	2.0	7.0	2.0	0	0	0	0	1	0	1	1	1	0	0	0	0	1
1	18.0	5	18.49	4.0	4	21.0	5	4.0	18.0	0	15.0	27.83	70.0	70.0	13.0	15.0	3.0	18.0	1	2002.0	1.0	1.0	0.0	3.0	5.0	5.0	10.0	7.0	7.0	7.0	13.0	5.0	13.0	2.0	0	0	1	0	0	0	0	1	0	0	0	1	0	0
2	12.0	5	16.99	2.0	4	3.0	8	7.0	30.0	0	33.0	22.77	67.0	67.0	11.0	4.0	5.0	27.0	0	2006.0	0.0	1.0	0.0	0.0	3.0	3.0	0.0	0.0	21.0	4.0	5.0	3.0	11.0	4.0	0	0	0	1	0	0	0	1	0	0	1	0	0	0
4	3.0	3	12.99	1.0	3	0.0	10	2.0	31.0	0	30.0	32.16	69.0	69.0	12.0	2.0	3.0	27.0	0	1977.0	0.0	1.0	1.0	2.0	7.0	7.0	2.0	4.0	9.0	10.0	15.0	7.0	12.0	4.0	0	1	0	0	0	1	0	1	0	1	0	0	1	0
6	2.0	3	7.69	0.0	1	18.0	9	3.0	26.0	0	51.0	17.49	75.0	75.0	12.0	3.0	0.0	23.0	0	2006.0	0.0	1.0	0.0	1.0	3.0	3.0	7.0	11.0	3.0	10.0	18.0	3.0	12.0	3.0	1	0	0	0	0	0	0	0	0	0	1	0	0	0

3.2 特征交互

3.3 特征编码

3.4 特征选择

# 相关性绝对值小于0.003的特征删除
data.corr()

	id	loanAmnt	term	interestRate	installment	grade	employmentTitle	employmentLength	annualIncome	issueDate	isDefault	postCode	dti	delinquency_2years	ficoRangeLow	ficoRangeHigh	openAcc	pubRec	pubRecBankruptcies	revolBal	revolUtil	totalAcc	initialListStatus	applicationType	earliesCreditLine	title	policyCode	n0	n1	n2	n3	n4	n5	n6	n7	n8	n9	n10	n11	n12	n13	n14	subGrade_2	subGrade_3	subGrade_4	subGrade_5	subGrade_6	subGrade_7	subGrade_8	subGrade_9	subGrade_10	subGrade_11	subGrade_12	subGrade_13	subGrade_14	subGrade_15	subGrade_16	subGrade_17	subGrade_18	subGrade_19	subGrade_20	subGrade_21	subGrade_22	subGrade_23	subGrade_24	subGrade_25	subGrade_26	subGrade_27	subGrade_28	subGrade_29	subGrade_30	subGrade_31	subGrade_32	subGrade_33	subGrade_34	subGrade_35	homeOwnership_1	homeOwnership_2	homeOwnership_3	homeOwnership_4	homeOwnership_5	verificationStatus_1	verificationStatus_2	purpose_1	purpose_2	purpose_3	purpose_4	purpose_5	purpose_6	purpose_7	purpose_8	purpose_9	purpose_10	purpose_11	purpose_12	purpose_13	regionCode_1	regionCode_2	regionCode_3	regionCode_4	regionCode_5	regionCode_6	regionCode_7	regionCode_8	regionCode_9	regionCode_10	regionCode_11	regionCode_12	regionCode_13	regionCode_14	regionCode_15	regionCode_16	regionCode_17	regionCode_18	regionCode_19	regionCode_20	regionCode_21	regionCode_22	regionCode_23	regionCode_24	regionCode_25	regionCode_26	regionCode_27	regionCode_28	regionCode_29	regionCode_30	regionCode_31	regionCode_32	regionCode_33	regionCode_34	regionCode_35	regionCode_36	regionCode_37	regionCode_38	regionCode_39	regionCode_40	regionCode_41	regionCode_42	regionCode_43	regionCode_44	regionCode_45	regionCode_46	regionCode_47	regionCode_48	regionCode_49	regionCode_50
id	1.000000	0.000423	-0.000832	0.001380	0.000935	0.001117	-0.000515	0.000522	0.001601	0.000813	0.000054	0.002119	-0.001419	0.000611	-0.000891	-0.000891	-0.002579	-0.000380	-0.001113	-0.002203	0.001751	-0.001191	0.001625	0.001088	-0.000648	-0.000831	NaN	0.001187	-0.001507	-0.001596	-0.001596	-0.001167	-0.000726	-0.001190	-0.002063	-0.001261	-0.001383	-0.002682	-0.000979	0.000919	-0.000399	0.000466	-0.003861	0.002312	-0.000586	0.001854	0.001601	-0.000564	-0.001624	-0.001781	-0.000133	-0.002010	0.000298	-0.000805	0.000330	0.000902	0.000199	0.001544	0.000008	0.001353	0.002652	-0.000004	0.001617	0.000757	-0.001725	-0.001176	0.000167	0.000584	-0.000396	0.001567	-0.001031	-0.001643	-0.000925	0.000106	0.000632	0.001376	0.001050	-0.000191	-0.000324	-0.001007	-0.000076	-0.001066	0.002051	0.001086	-0.000949	0.000830	-0.001223	0.001617	0.001010	0.000869	0.000494	-0.000550	-0.001030	-0.000046	0.000530	-0.001829	-0.000512	0.000132	-0.002777	0.000837	0.000781	0.001360	0.000291	0.003432	0.000811	0.000275	-0.000130	-0.002148	-0.000562	-0.001097	0.000369	-0.000868	0.000086	0.000242	-0.001352	0.000190	0.002099	-0.001709	0.000090	-0.002066	-0.001333	0.001156	0.000646	-0.001189	0.001397	0.000154	-0.000731	-0.003658	-0.002803	-0.000918	0.000107	0.000119	0.001542	-0.000442	-0.001099	-0.000408	0.000445	0.000052	0.000535	-0.002393	0.001781	0.000439	0.001373	-0.000621	-0.001178	NaN
loanAmnt	0.000423	1.000000	0.409910	0.118841	0.944344	0.120627	-0.009845	0.065608	0.467275	-0.003040	0.059117	-0.021573	0.027657	0.011335	0.112145	0.112145	0.183657	-0.085582	-0.095282	0.441242	0.115935	0.224888	-0.064040	0.063317	-0.163749	-0.019045	NaN	-0.039067	0.185727	0.139590	0.139590	0.205452	0.199770	0.096398	0.158380	0.173734	0.140495	0.179692	-0.000524	0.001787	-0.009348	-0.039259	-0.015960	-0.013134	-0.007710	-0.006264	-0.029218	-0.026768	-0.023572	-0.028199	-0.041442	-0.023762	-0.015886	0.001133	0.016117	0.011130	0.001577	-0.000291	0.009578	0.026116	0.028586	0.033834	0.038488	0.041068	0.045713	0.048670	0.036593	0.039976	0.030056	0.029536	0.028785	0.025081	0.018754	0.020721	0.016866	0.016305	-0.164712	-0.026631	-0.002439	0.001072	-0.000577	0.030942	0.155035	0.008906	-0.018050	-0.044906	0.029625	-0.137541	0.002530	-0.077588	-0.066592	-0.067713	-0.065535	-0.014204	-0.015790	-0.003060	-0.003828	0.005890	0.006220	-0.003724	-0.003024	-0.005583	-0.004681	0.006280	0.017927	-0.010561	-0.005282	0.012463	-0.006248	0.027367	-0.002174	-0.000754	-0.004629	0.006900	-0.014687	-0.000519	-0.024489	0.003046	-0.012151	-0.007140	0.004722	0.016062	-0.005721	0.006328	-0.004399	0.017766	-0.000734	-0.001726	0.000768	-0.005510	-0.011785	0.001404	-0.007696	-0.007866	0.000001	-0.004725	0.005218	-0.000192	-0.010283	-0.003038	0.002874	0.000126	-0.001679	0.013122	-0.001992	NaN
term	-0.000832	0.409910	1.000000	0.415437	0.162911	0.423747	0.013595	0.043317	0.106556	-0.038790	0.172367	0.011870	0.064007	-0.003635	0.010680	0.010680	0.077523	-0.019022	-0.011780	0.133292	0.066051	0.117111	-0.093668	0.039031	-0.055153	0.006475	NaN	-0.015322	0.042665	0.050506	0.050506	0.047499	0.058447	0.084320	0.051615	0.064934	0.050742	0.079613	0.000024	-0.000450	-0.005690	0.021253	-0.092775	-0.093894	-0.102663	-0.095506	-0.079971	-0.077644	-0.065284	-0.055868	-0.074103	-0.026473	-0.005394	0.033247	0.062081	0.057930	0.043461	0.043473	0.055638	0.083301	0.090898	0.096988	0.103288	0.109382	0.116059	0.117178	0.093415	0.089779	0.071815	0.071525	0.060260	0.053811	0.044648	0.035705	0.028910	0.027364	-0.099142	-0.020997	-0.001623	-0.001340	-0.000943	0.033459	0.088802	0.002053	0.004970	-0.013546	-0.038622	-0.050243	0.005194	-0.034611	-0.018643	-0.023816	-0.027348	-0.006035	-0.003893	0.002572	0.000277	-0.007993	0.003824	0.003904	0.004104	0.001358	0.011006	-0.021286	0.012225	-0.001926	0.008378	0.009153	-0.010520	-0.005189	0.001348	-0.000885	0.008042	-0.000223	0.007868	0.003341	-0.012818	0.005266	0.000807	0.002217	-0.003946	0.003221	0.002844	0.001443	-0.001787	0.001891	0.003290	0.005010	-0.000667	-0.001398	-0.003144	-0.001115	0.002452	-0.007542	0.009668	0.000364	0.004289	0.003000	0.006001	-0.000352	0.008120	-0.002490	-0.001081	0.000969	0.000010	NaN
interestRate	0.001380	0.118841	0.415437	1.000000	0.120640	0.951939	0.072586	-0.000166	-0.128969	-0.038144	0.253111	0.008645	0.178430	0.041478	-0.397102	-0.397102	-0.026849	0.052064	0.048077	-0.029575	0.240812	-0.071632	0.129388	0.029456	0.107659	0.016223	NaN	0.042391	0.013137	0.082619	0.082619	-0.066004	-0.093405	-0.015900	-0.020434	-0.069891	0.081010	-0.030676	0.003671	0.012847	0.026751	0.190942	-0.243300	-0.225212	-0.249667	-0.245046	-0.221699	-0.173199	-0.134540	-0.093975	-0.063343	-0.020932	0.018093	0.049204	0.086881	0.125377	0.144427	0.169726	0.184474	0.198299	0.208694	0.197218	0.206423	0.209187	0.214386	0.233260	0.170868	0.158553	0.152411	0.141718	0.124198	0.107149	0.085428	0.060953	0.047413	0.044686	0.063943	0.007200	0.000583	0.002810	0.002654	0.014635	0.208091	0.064548	-0.019210	-0.012516	-0.168315	0.074356	0.036263	0.009930	-0.022643	0.020322	0.036275	0.012358	0.010037	0.000691	0.001799	-0.006851	0.000489	-0.004114	0.002838	0.000529	0.005557	-0.007345	0.001095	-0.005518	0.001007	0.002319	0.011774	-0.007952	-0.002154	-0.001417	0.011875	0.002459	0.002773	0.000049	0.005241	-0.005651	0.004637	-0.001894	-0.006674	-0.015917	-0.004497	0.009403	-0.003376	-0.003325	0.000832	0.006044	-0.005375	0.001907	-0.003610	-0.003600	0.007047	0.003819	0.000419	-0.001031	0.000135	0.003551	0.001716	0.002648	-0.000331	0.001871	-0.002838	0.002601	0.002269	NaN
installment	0.000935	0.944344	0.162911	0.120640	1.000000	0.115138	-0.005217	0.055356	0.440590	-0.000578	0.041954	-0.024978	0.034484	0.018761	0.063176	0.063176	0.169524	-0.076187	-0.089488	0.418453	0.133743	0.193653	-0.019355	0.055460	-0.142758	-0.019282	NaN	-0.031052	0.188863	0.146766	0.146766	0.197947	0.184984	0.073793	0.153171	0.158454	0.147354	0.164535	-0.000376	0.003462	-0.004688	-0.024037	-0.012639	-0.008244	-0.003798	-0.003254	-0.031604	-0.027494	-0.027593	-0.030215	-0.035670	-0.018203	-0.011481	-0.004921	0.006126	0.004602	0.002066	0.005157	0.013816	0.025210	0.027891	0.030986	0.034989	0.037623	0.042111	0.048661	0.032765	0.038853	0.032514	0.031743	0.032160	0.027272	0.020309	0.021406	0.016787	0.016617	-0.134317	-0.020577	-0.002310	0.001835	-0.000081	0.023925	0.159702	0.017272	-0.024932	-0.046447	0.024458	-0.129208	0.004799	-0.075295	-0.066935	-0.065726	-0.061881	-0.012486	-0.014627	-0.003316	-0.003710	0.008499	0.006028	-0.005617	-0.003991	-0.006158	-0.008413	0.012794	0.014852	-0.011654	-0.008282	0.009407	-0.001891	0.029017	-0.002955	-0.001005	-0.006236	0.007959	-0.018127	-0.001813	-0.021365	0.000745	-0.012224	-0.008418	0.005480	0.016269	-0.006885	0.007615	-0.004297	0.017403	-0.001918	-0.002410	0.000553	-0.005730	-0.011770	0.001309	-0.008045	-0.005362	-0.003144	-0.004970	0.004123	-0.000890	-0.012445	-0.002893	0.000689	0.001379	-0.002107	0.014514	-0.002434	NaN
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
regionCode_46	0.000439	0.000126	-0.002490	0.001871	0.001379	0.002765	-0.008369	-0.011931	-0.004625	0.024831	-0.000336	0.058462	0.009459	0.003612	-0.002016	-0.002016	-0.003194	0.001476	-0.002488	-0.004314	0.002978	-0.001576	-0.013439	0.009129	0.011684	-0.005360	NaN	-0.000041	-0.002700	-0.004665	-0.004665	-0.004031	-0.008001	0.009072	-0.007103	-0.008401	-0.005226	-0.003726	0.001148	0.001783	0.000147	0.002509	-0.002335	-0.002496	-0.001349	-0.003406	0.001946	-0.000224	-0.001426	-0.000537	0.002037	-0.000117	-0.000140	0.000594	0.000096	0.005490	0.001900	0.000297	0.002810	-0.000436	-0.001953	0.000466	-0.000527	0.001300	-0.002110	-0.000180	0.000703	0.000282	-0.000673	-0.000245	-0.000626	-0.001348	-0.001072	-0.000790	-0.000618	-0.000568	-0.001450	0.002148	-0.000553	-0.000206	-0.000190	0.001963	0.001363	-0.000022	-0.002358	-0.001070	0.002238	0.000232	-0.001839	0.000393	0.000282	0.001306	0.000258	0.002608	-0.001346	-0.000119	-0.001614	-0.007011	-0.006412	-0.004771	-0.003901	-0.002357	-0.006112	-0.014704	-0.006070	-0.005696	-0.004578	-0.005392	-0.010477	-0.010645	-0.003315	-0.002605	-0.003985	-0.005350	-0.006564	-0.003865	-0.009814	-0.005398	-0.005811	-0.004546	-0.001762	-0.005437	-0.004141	-0.002559	-0.001590	-0.006658	-0.001902	-0.004436	-0.002476	-0.001846	-0.003979	-0.004241	-0.003088	-0.004423	-0.002117	-0.001948	-0.001703	-0.003432	-0.003531	-0.002525	-0.003082	1.000000	-0.001388	-0.001764	-0.000718	NaN
regionCode_47	0.001373	-0.001679	-0.001081	-0.002838	-0.002107	-0.002504	-0.007356	0.001941	-0.007198	0.030269	-0.005557	0.060215	0.007797	0.000471	0.001091	0.001091	-0.003704	0.003830	0.002131	0.000469	0.000210	0.000105	-0.013378	0.009576	0.002036	-0.005905	NaN	-0.000110	-0.000971	-0.004866	-0.004866	-0.000450	-0.001806	0.004435	-0.006316	-0.004431	-0.005377	-0.004324	-0.000898	-0.000308	-0.000536	-0.003055	-0.000755	-0.000961	-0.001134	0.001827	0.001583	0.000080	-0.001527	0.002260	0.002075	-0.001430	-0.000158	0.001319	0.001773	-0.001070	0.001429	-0.000536	0.001861	0.001303	-0.001691	-0.001744	-0.001350	-0.001749	-0.000725	-0.001673	-0.000952	-0.000817	-0.001704	-0.000594	-0.000881	-0.001484	-0.001180	-0.000870	-0.000681	-0.000625	-0.009648	0.005573	-0.000609	-0.000226	-0.000210	0.001542	-0.001215	-0.001125	-0.001645	-0.000254	0.003326	-0.001654	-0.001564	0.001295	0.001141	-0.001868	-0.002305	0.000595	-0.001483	-0.000131	-0.001778	-0.007724	-0.007063	-0.005255	-0.004297	-0.002596	-0.006733	-0.016198	-0.006686	-0.006275	-0.005044	-0.005940	-0.011541	-0.011727	-0.003651	-0.002869	-0.004390	-0.005894	-0.007231	-0.004257	-0.010811	-0.005946	-0.006402	-0.005008	-0.001941	-0.005990	-0.004562	-0.002819	-0.001752	-0.007335	-0.002096	-0.004887	-0.002727	-0.002033	-0.004384	-0.004672	-0.003402	-0.004872	-0.002332	-0.002145	-0.001876	-0.003781	-0.003890	-0.002782	-0.003395	-0.001388	1.000000	-0.001943	-0.000791	NaN
regionCode_48	-0.000621	0.013122	0.000969	0.002601	0.014514	0.002697	0.005839	0.001035	0.005677	-0.006643	0.000386	0.074851	0.003538	-0.000505	-0.000066	-0.000066	-0.008804	-0.010687	-0.009029	0.012497	0.016935	-0.002152	0.003535	-0.000401	-0.000109	0.002563	NaN	-0.003413	-0.004071	-0.006308	-0.006308	-0.006013	-0.003305	0.003450	-0.011382	-0.006447	-0.005964	-0.008729	-0.001141	-0.001788	-0.000788	-0.007267	-0.002167	0.000319	0.001117	0.000323	0.000613	-0.001174	-0.002295	-0.001939	0.001481	-0.000007	-0.001138	0.000008	-0.001269	-0.000826	0.001418	0.001531	-0.000812	0.002184	-0.000215	0.001059	0.000417	0.000593	0.001657	-0.000144	-0.000153	0.003352	-0.000897	-0.000291	-0.000136	0.000664	0.002774	0.001791	0.000984	0.001222	-0.002532	0.001040	-0.000774	-0.000288	0.005743	0.000934	0.001735	0.001361	0.000426	-0.000243	-0.001842	-0.003017	0.002071	0.003414	-0.001371	0.000768	0.001233	-0.000015	0.000669	-0.000166	-0.002259	-0.009814	-0.008975	-0.006678	-0.005460	-0.003299	-0.008555	-0.020582	-0.008496	-0.007973	-0.006408	-0.007547	-0.014664	-0.014901	-0.004640	-0.003646	-0.005578	-0.007488	-0.009188	-0.005409	-0.013736	-0.007556	-0.008134	-0.006364	-0.002466	-0.007610	-0.005797	-0.003582	-0.002226	-0.009320	-0.002663	-0.006209	-0.003465	-0.002583	-0.005570	-0.005937	-0.004323	-0.006191	-0.002963	-0.002726	-0.002384	-0.004804	-0.004943	-0.003534	-0.004314	-0.001764	-0.001943	1.000000	-0.001005	NaN
regionCode_49	-0.001178	-0.001992	0.000010	0.002269	-0.002434	0.000968	-0.003484	0.000365	-0.006709	0.020769	-0.002056	0.059028	0.004764	-0.000256	-0.001166	-0.001166	-0.001209	0.011928	0.004670	-0.003665	-0.000948	-0.001023	-0.006241	0.015515	0.005007	-0.003054	NaN	0.003467	-0.004131	-0.001632	-0.001632	-0.004665	-0.007655	0.002124	-0.001115	-0.003804	-0.001842	-0.001661	-0.000464	0.000555	0.001177	0.001127	0.000018	-0.000573	-0.003264	-0.003465	-0.000607	0.000914	-0.000261	0.001553	0.002841	-0.000248	0.000493	0.002180	0.000185	-0.000020	0.000113	-0.001948	0.001733	0.000162	0.000193	-0.000263	0.001869	-0.000318	0.000810	-0.001332	-0.001581	-0.000191	-0.001230	-0.001093	0.000817	-0.000768	-0.000610	-0.000450	-0.000352	-0.000323	-0.007148	-0.001560	-0.000315	-0.000117	-0.000108	-0.000953	0.002061	-0.001269	-0.000138	-0.001287	-0.002003	0.001516	-0.001416	-0.000695	-0.000552	0.001566	0.000195	-0.000518	-0.000767	-0.000068	-0.000920	-0.003995	-0.003653	-0.002718	-0.002222	-0.001343	-0.003482	-0.008377	-0.003458	-0.003245	-0.002608	-0.003072	-0.005969	-0.006065	-0.001888	-0.001484	-0.002271	-0.003048	-0.003740	-0.002202	-0.005591	-0.003075	-0.003311	-0.002590	-0.001004	-0.003098	-0.002359	-0.001458	-0.000906	-0.003794	-0.001084	-0.002527	-0.001411	-0.001052	-0.002267	-0.002416	-0.001760	-0.002520	-0.001206	-0.001110	-0.000970	-0.001955	-0.002012	-0.001439	-0.001756	-0.000718	-0.000791	-0.001005	1.000000	NaN
regionCode_50	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN

146 rows × 146 columns

corr = data.corr().unstack()['isDefault'].sort_values(ascending=False)
drop_fea = corr[abs(corr)<0.003].index
data.drop(drop_fea,axis=1,inplace=True)

data.head()

	loanAmnt	term	interestRate	installment	grade	employmentTitle	annualIncome	issueDate	isDefault	postCode	dti	ficoRangeLow	ficoRangeHigh	openAcc	revolBal	revolUtil	totalAcc	initialListStatus	earliesCreditLine	title	policyCode	n0	n1	n2	n3	n4	n5	n6	n7	n8	n9	n10	n14	subGrade_4	subGrade_12	subGrade_17	subGrade_18	subGrade_22	homeOwnership_1	homeOwnership_2	verificationStatus_2	purpose_1	purpose_10	regionCode_14	regionCode_18	regionCode_21	regionCode_32
0	35.0	5	19.52	9.0	5	0.0	11.0	25.0	1	13.0	17.05	73.0	73.0	7.0	24.0	4.0	27.0	0	2001.0	0.0	1.0	0.0	2.0	2.0	2.0	4.0	9.0	8.0	4.0	12.0	2.0	7.0	2.0	0	0	0	0	1	0	1	1	1	0	0	0	0	1
1	18.0	5	18.49	4.0	4	21.0	4.0	18.0	0	15.0	27.83	70.0	70.0	13.0	15.0	3.0	18.0	1	2002.0	1.0	1.0	0.0	3.0	5.0	5.0	10.0	7.0	7.0	7.0	13.0	5.0	13.0	2.0	0	0	1	0	0	0	0	1	0	0	0	1	0	0
2	12.0	5	16.99	2.0	4	3.0	7.0	30.0	0	33.0	22.77	67.0	67.0	11.0	4.0	5.0	27.0	0	2006.0	0.0	1.0	0.0	0.0	3.0	3.0	0.0	0.0	21.0	4.0	5.0	3.0	11.0	4.0	0	0	0	1	0	0	0	1	0	0	1	0	0	0
4	3.0	3	12.99	1.0	3	0.0	2.0	31.0	0	30.0	32.16	69.0	69.0	12.0	2.0	3.0	27.0	0	1977.0	0.0	1.0	1.0	2.0	7.0	7.0	2.0	4.0	9.0	10.0	15.0	7.0	12.0	4.0	0	1	0	0	0	1	0	1	0	1	0	0	1	0
6	2.0	3	7.69	0.0	1	18.0	3.0	26.0	0	51.0	17.49	75.0	75.0	12.0	3.0	0.0	23.0	0	2006.0	0.0	1.0	0.0	1.0	3.0	3.0	7.0	11.0	3.0	10.0	18.0	3.0	12.0	3.0	1	0	0	0	0	0	0	0	0	0	1	0	0	0

四.建型与调参

4.1 划分训练集/测试集/验证集

x_train = data.drop(['isDefault'],axis=1)
y_train = data['isDefault']x, val_x, y, val_y = train_test_split(x_train,y_train,test_size=0.25,random_state=1)

4.2 选择模型

4.2.1 逻辑回归模型

lr = LogisticRegression()
scores = cross_val_score(lr,x_train,y_train,cv=5,scoring='accuracy')
print('逻辑回归5折交叉训练准确率为：',np.mean(scores))

逻辑回归5折交叉训练准确率为： 0.803408265122779

4.2.2 随机森林模型

rfc = RandomForestClassifier(n_estimators=250,max_depth=10)
rfc.fit(x,y)
print('随机森林准确率为：',rfc.score(val_x,val_y))

随机森林准确率为： 0.8049472312370634

4.3 网格搜索进行超参数调优

param_grid = {'n_estimators':[100,150,200,300],'max_depth':[5,10,15]}
r = RandomForestClassifier()
emstimator = GridSearchCV(r,param_grid=param_grid,cv=5)
emstimator.fit(x_train,y_train)
emstimator.best_score_

五、总结

金融风控的实际项目多涉及到信用评分，因此需要模型特征具有较好的可解释性，所以目前在实际项目中多还是以逻辑回归作为基础模型。

如果想获得更好的结果，可以使用集成算法进行建模，但解释性不强。

35. 贷款违约预测相关推荐

【算法竞赛学习】金融风控之贷款违约预测-建模与调参
Task4 建模与调参此部分为零基础入门金融风控的 Task4 建模调参部分,带你来了解各种模型以及模型的评价和调参策略,欢迎大家后续多多交流. 赛题:零基础入门数据挖掘 - 零基础入门金融风控之贷 ...
数据竞赛入门-金融风控（贷款违约预测）四、建模与调参
前言本次活动为datawhale与天池联合举办,为金融风控之贷款违约预测挑战赛(入门) 比赛地址:https://tianchi.aliyun.com/competition/entrance/53 ...
数据竞赛入门-金融风控（贷款违约预测）三、特征工程
前言本次活动为datawhale与天池联合举办,为金融风控之贷款违约预测挑战赛(入门) 比赛地址:https://tianchi.aliyun.com/competition/entrance/53 ...
基于机器学习与深度学习的金融风控贷款违约预测
基于机器学习与深度学习的金融风控贷款违约预测目录一.赛题分析 1. 任务分析 2. 数据属性 3. 评价指标 4. 问题归类 5. 整体思路二.数据可视化分析 1. 总体数据分析 2. 数值型数 ...
「机器学习」天池比赛：金融风控贷款违约预测
一.前言 1.1 赛题背景赛题以金融风控中的个人信贷为背景,要求选手根据贷款申请人的数据信息预测其是否有违约的可能,以此判断是否通过此项贷款,这是一个典型的分类问题. 任务:预测用户贷款是否违约比 ...
数据挖掘机器学习[六]---项目实战金融风控之贷款违约预测
相关文章: 特征工程详解及实战项目[参考] 数据挖掘---汽车车交易价格预测[一](测评指标:EDA) 数据挖掘机器学习---汽车交易价格预测详细版本[二]{EDA-数据探索性分析} 数据挖掘机器学习 ...
1.天池金融风控-贷款违约预测新人赛之预备知识
比赛链接:金融风控-贷款违约预测因为这是一个金融风控专题的数据挖掘实战,在开始之前先引入一些预备知识. 1.预备知识 1.1预测指标本次竞赛用AUC作为评价指标,AUC为ROC曲线下与坐标轴围成的 ...
入门金融风控【贷款违约预测】
入门金融风控[贷款违约预测] 赛题以金融风控中的个人信贷为背景,要求选手根据贷款申请人的数据信息预测其是否有违约的可能,以此判断是否通过此项贷款,这是一个典型的分类问题.通过这道赛题来引导大家了解金融 ...
金融风控-贷款违约预测学习笔记(Part3：特征工程)
金融风控-贷款违约预测学习笔记(Part3:特征工程) 1.特征预处理 1.1 处理类别型特征和数值型特征 1.2 缺失值填充 1.3 时间格式处理 1.4 将对象类型特征转换到数值 1.5 类别特征 ...

35. 贷款违约预测