阿里云金融风控-贷款违约预测建模

直接附上代码

# -*- coding: utf-8 -*-
"""
Created on Sat Jan 16 15:18:33 2021

@author: Administrator
"""

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
import datetime
import warnings
warnings.filterwarnings('ignore')

#%%导入数据
data_train = pd.read_csv('D:/python_home/阿里云金融风控-贷款违约预测/train.csv')
data_test_a = pd.read_csv('D:/python_home/阿里云金融风控-贷款违约预测/testA.csv')


#%%基本的数据描述
data_train.shape,data_test_a.shape  #((800000, 47), (200000, 46))

data_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')
'''

data_test_a.columns

'''
Index(['id', 'loanAmnt', 'term', 'interestRate', 'installment', 'grade',
       'subGrade', 'employmentTitle', 'employmentLength', 'homeOwnership',
       'annualIncome', 'verificationStatus', 'issueDate', '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')
'''

#%%看一下变量的取值的个数，用来划分数值型还是类别型
for i in list(data_train.columns):
    print(i, data_train[i].nunique())
'''
id 800000
loanAmnt 1540  贷款金额
term 2
interestRate 641
installment 72360
grade 7
subGrade 35
employmentTitle 248683
employmentLength 11
homeOwnership 6
annualIncome 44926
verificationStatus 3
issueDate 139
isDefault 2
purpose 14
postCode 932
regionCode 51
dti 6321
delinquency_2years 30
ficoRangeLow 39
ficoRangeHigh 39
openAcc 75
pubRec 32
pubRecBankruptcies 11
revolBal 71116
revolUtil 1286
totalAcc 134
initialListStatus 2
applicationType 2
earliesCreditLine 720
title 39644
policyCode 1
n0 39
n1 33
n2 50
n3 50
n4 46
n5 65
n6 107
n7 70
n8 102
n9 44
n10 76
n11 5
n12 5
n13 28
n14 31
'''

cate_col = ['term', 'grade',
       'subGrade', 'employmentLength', 'homeOwnership',
        'verificationStatus',  'isDefault',
       'purpose', 
       'pubRecBankruptcies', 
       'initialListStatus', 'applicationType', 
       'policyCode','n11', 'n12']

num_col = [i for i in list(data_train.columns)[1:] if i not in cate_col]

data_train[cate_col].nunique()


#%%类别变量的iv值计算
import pycard as pc
cate_iv_woedf = pc.WoeDf()
clf = pc.NumBin()
for i in cate_col:
    cate_iv_woedf.append(pc.cross_woe(data_train[i] ,data_train.isDefault))
cate_iv_woedf.to_excel('tmp11')

cate_use_col = ['term','grade','verificationStatus']

#%%数值型变量的iv值计算
num_col.remove('issueDate')
num_col.remove('earliesCreditLine')
#上面这两个是时间日期的东西，后面再做处理吧

num_iv_woedf = pc.WoeDf()
clf = pc.NumBin()
for i in num_col:
    clf.fit(data_train[i] ,data_train.isDefault)
    clf.generate_transform_fun()
    num_iv_woedf.append(clf.woe_df_)
num_iv_woedf.to_excel('tmp12')

from numpy import *
data_train['loanAmnt_bin'] = pd.cut(data_train.loanAmnt,bins=[-inf, 3512.5, 9012.5,  10012.5, 11987.5, 15012.5, 28012.5, inf])

#interestRate
data_train['interestRate_bin'] = pd.cut(data_train.interestRate,bins=[-inf, 7.885, 9.73, 11.415, 13.175, 15.975, 17.785, 21.985, inf])

#annualIncome
data_train['annualIncome_bin'] = pd.cut(data_train.annualIncome,bins=[-inf, 37001.5996, 45670.5, 60995.5, 70017.5, 86462.0, 100670.5, 160030.0, inf])

#dti,先用均值填充，再分
data_train['dti'] = data_train['dti'].fillna(data_train['dti'].mean())
data_train['dti_bin'] = pd.cut(data_train.dti,bins=[-inf, 10.745, 14.845, 18.255, 21.745, 25.325, 30.195, 33.225, inf])

#ficoRangeLow
data_train['ficoRangeLow_bin'] = pd.cut(data_train.ficoRangeLow,bins=[-inf, 667.5, 682.5, 692.5, 702.5, 717.5, 732.5, 767.5, inf])

#revolUtil，均值填充，再分
data_train['revolUtil'] = data_train['revolUtil'].fillna(data_train['revolUtil'].mean())
data_train['revolUtil_bin'] = pd.cut(data_train.revolUtil,bins=[-inf, 19.75, 29.35, 38.55, 47.95, 56.55, 86.85, inf])

#n14 空值作为一列，
data_train['n14_bin'] = pd.cut(data_train.n14,bins=[-inf, 0.5, 1.5, 2.5, 3.5, 4.5, 6.5, inf])

woe_col = [i for i in ['term','grade','verificationStatus']+list(data_train.columns)[-7:]]

#%%


cate_iv_woedf = pc.WoeDf()
clf = pc.NumBin()
for i in woe_col:
    cate_iv_woedf.append(pc.cross_woe(data_train[i] ,data_train.isDefault))
cate_iv_woedf.to_excel('tmp11')

data_train.grade[data_train.grade =='G'] = 'F'

#%%woe转换
pc.obj_info(cate_iv_woedf)

cate_iv_woedf.bin2woe(data_train,woe_col)

model_col = [i for i in ['id', 'isDefault']+list(data_train.columns)[-10:]]

data_train[model_col].isnull().sum()
data_train[model_col].info()
model_data = data_train[model_col]
model_data = model_data.astype(float)
model_data.n14_woe[model_data.n14_woe.isnull()]=0.34984133

#%%建模
import pandas as pd
import matplotlib.pyplot as plt #导入图像库
import matplotlib
import seaborn as sns
import statsmodels.api as sm
from sklearn.metrics import roc_curve, auc
from sklearn.model_selection import train_test_split

X = model_data[['_woe',
 'g_woe',
 'verificationSt_woe',
 'loanAmnt_woe',
 'interestRate_woe',
 'annualIncome_woe',
 'dti_woe',
 'ficoRangeLow_woe',
 'revolUtil_woe',
 'n14_woe']]
Y = model_data['isDefault']


x_train,x_test,y_train,y_test=train_test_split(X,Y,test_size=0.3,random_state=0)

#(10127, 44)

X1=sm.add_constant(x_train)   #在X前加上一列常数1，方便做带截距项的回归
logit=sm.Logit(y_train.astype(float),X1.astype(float))
result=logit.fit()
result.summary()
result.params



X3 = sm.add_constant(x_test)
resu = result.predict(X3.astype(float))
fpr, tpr, threshold = roc_curve(y_test, resu)
rocauc = auc(fpr, tpr)
plt.plot(fpr, tpr, 'b', label='AUC = %0.2f' % rocauc)
plt.legend(loc='lower right')
plt.plot([0, 1], [0, 1], 'r--')
plt.xlim([0, 1])
plt.ylim([0, 1])
plt.ylabel('真正率')
plt.xlabel('假正率')
plt.show()


resu_1 = result.predict(X1.astype(float))
fpr, tpr, threshold = roc_curve(y_train, resu_1)
rocauc = auc(fpr, tpr)
plt.plot(fpr, tpr, 'b', label='AUC = %0.2f' % rocauc)
plt.legend(loc='lower right')
plt.plot([0, 1], [0, 1], 'r--')
plt.xlim([0, 1])
plt.ylim([0, 1])
plt.ylabel('真正率')
plt.xlabel('假正率')
plt.show()

 最后结果是 0.7026，比上次的0.6829 多了约2%，目前排名榜上最高分是0.7492，距离目标还差5%左右

看了看别人写的代码，效果能达到7.348，现附上链接：https://blog.csdn.net/qq_44694861/article/details/109753004?spm=5176.12282029.0.0.209a4288OlvFjo

代码如下：

# -*- coding: utf-8 -*-
"""
Created on Tue Feb  9 10:04:26 2021

@author: Administrator
"""

#%%
import pandas as pd
import datetime
import warnings
warnings.filterwarnings('ignore')
from sklearn.model_selection import StratifiedKFold
#warnings.filterwarnings('ignore')
#%matplotlib inline
from sklearn.metrics import roc_auc_score
## 数据降维处理的
from sklearn.model_selection import train_test_split  
from catboost import CatBoostClassifier

#pip3 install --user  catboost -i https://pypi.tuna.tsinghua.edu.cn/simple/

#%%
train = pd.read_csv('D:/python_home/阿里云金融风控-贷款违约预测/train.csv')
testA = pd.read_csv('D:/python_home/阿里云金融风控-贷款违约预测/testA.csv')


#%%
numerical_fea = list(train.select_dtypes(exclude=['object']).columns)
numerical_fea.remove('isDefault')
train[numerical_fea] = train[numerical_fea].fillna(train[numerical_fea].median())
testA[numerical_fea] = testA[numerical_fea].fillna(testA[numerical_fea].median())
#issueDate
for data in [train]:
    data['issueDate'] = pd.to_datetime(data['issueDate'],format='%Y-%m-%d')
    data['grade'] = data['grade'].map({'A':1,'B':2,'C':3,'D':4,'E':5,'F':6,'G':7})
    data['employmentLength'] = data['employmentLength'].map({'1 year':1,'2 years':2,'3 years':3,'4 years':4,'5 years':5,'6 years':6,'7 years':7,'8 years':8,'9 years':9,'10+ years':10,'< 1 year':0})
    data['subGrade'] = data['subGrade'].map({'E2':1,'D2':2,'D3':3,'A4':4,'C2':5,'A5':6,'C3':7,'B4':8,'B5':9,'E5':10,
        'D4':11,'B3':12,'B2':13,'D1':14,'E1':15,'C5':16,'C1':17,'A2':18,'A3':19,'B1':20,
        'E3':21,'F1':22,'C4':23,'A1':24,'D5':25,'F2':26,'E4':27,'F3':28,'G2':29,'F5':30,
        'G3':31,'G1':32,'F4':33,'G4':34,'G5':35})
    data['earliesCreditLine'] = data['earliesCreditLine'].apply(lambda s: int(s[-4:]))
  #  data['n15']=data['n8']*data['n10']
    
for data in [testA]:
    data['issueDate'] = pd.to_datetime(data['issueDate'],format='%Y-%m-%d')
    data['grade'] = data['grade'].map({'A':1,'B':2,'C':3,'D':4,'E':5,'F':6,'G':7})
    data['employmentLength'] = data['employmentLength'].map({'1 year':1,'2 years':2,'3 years':3,'4 years':4,'5 years':5,'6 years':6,'7 years':7,'8 years':8,'9 years':9,'10+ years':10,'< 1 year':0})
    data['subGrade'] = data['subGrade'].map({'E2':1,'D2':2,'D3':3,'A4':4,'C2':5,'A5':6,'C3':7,'B4':8,'B5':9,'E5':10,
        'D4':11,'B3':12,'B2':13,'D1':14,'E1':15,'C5':16,'C1':17,'A2':18,'A3':19,'B1':20,
        'E3':21,'F1':22,'C4':23,'A1':24,'D5':25,'F2':26,'E4':27,'F3':28,'G2':29,'F5':30,
        'G3':31,'G1':32,'F4':33,'G4':34,'G5':35})
    data['earliesCreditLine'] = data['earliesCreditLine'].apply(lambda s: int(s[-4:]))

print("数据预处理完成!")  

#%%
sub=testA[['id']].copy()
sub['isDefault']=0
testA=testA.drop(['id','issueDate'],axis=1)
data_x=train.drop(['isDefault','id','issueDate'],axis=1)
data_y=train[['isDefault']].copy()
x, val_x, y, val_y = train_test_split(  
    data_x,  
    data_y,  
    test_size=0.25,  
    random_state=1,  
    stratify=data_y
)  

col=['grade','subGrade','employmentTitle','homeOwnership','verificationStatus','purpose','postCode','regionCode',
     'initialListStatus','applicationType','policyCode']
for i in data_x.columns:
    if i in col:
        data_x[i] = data_x[i].astype('str')
for i in testA.columns:
    if i in col:
        testA[i] = testA[i].astype('str')

#%%
model=CatBoostClassifier(
            loss_function="Logloss",
            eval_metric="AUC",
            task_type="CPU",
            learning_rate=0.1,
            iterations=500,
            random_seed=2020,
            od_type="Iter",
            depth=7)

answers = []
mean_score = 0
n_folds = 5
sk = StratifiedKFold(n_splits=n_folds, shuffle=True, random_state=2019)
for train, test in sk.split(data_x, data_y):
    x_train = data_x.iloc[train]
    y_train = data_y.iloc[train]
    x_test = data_x.iloc[test]
    y_test = data_y.iloc[test]
    clf = model.fit(x_train,y_train, eval_set=(x_test,y_test),verbose=500,cat_features=col)
    yy_pred_valid=clf.predict(x_test)
    print('cat验证的auc:{}'.format(roc_auc_score(y_test, yy_pred_valid)))
    mean_score += roc_auc_score(y_test, yy_pred_valid) / n_folds
    y_pred_valid = clf.predict(testA,prediction_type='Probability')[:,-1]
    answers.append(y_pred_valid)
print('mean valAuc:{}'.format(mean_score))

#%%
cat_pre=sum(answers)/n_folds
sub['isDefault']=cat_pre
sub.to_csv('金融预测.csv',index=False)

注意事项：

1.catboost只能识别字符类型和数值类型的数据

2.代码需要很长的时间去跑