kaggle_GiveMeSomeCredit_建模比赛

最新推荐文章于 2024-04-23 12:14:13 发布
最新推荐文章于 2024-04-23 12:14:13 发布
阅读量996 收藏 13
点赞数
分类专栏： 机器学习
版权声明：本文为博主原创文章，遵循 CC 4.0 BY-SA 版权协议，转载请附上原文出处链接和本声明。
本文链接：https://blog.csdn.net/lvhuike/article/details/113939718
版权
本文探讨了使用Logistic回归和随机森林算法在信用卡违约预测竞赛中的应用，通过特征工程、上采样和参数调优，优化了模型性能，最终达到privatescore0.86699和publicscore0.86101的成绩。
摘要由CSDN通过智能技术生成

kaggle提交结果：
在这里插入图片描述
我的private score 0.86699, public score 0.86101,
榜单第一名private score 0.86955, public score 0.86390.

Importing libraries

import pandas as pd
import numpy as np
import os

from sklearn.ensemble import RandomForestClassifier
from sklearn.ensemble import GradientBoostingClassifier
import imblearn
from imblearn.over_sampling import SMOTE
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
from sklearn.model_selection import GridSearchCV
from sklearn.model_selection import cross_val_score

import matplotlib.pyplot as plt
import seaborn as sns

import scikitplot as skplt
import scipy

from sklearn.model_selection import GridSearchCV, RandomizedSearchCV
from sklearn.metrics import roc_curve, roc_auc_score, confusion_matrix as cm, auc, roc_curve

Reading the data

data = pd.read_csv('cs-training.csv', index_col=0)

test_data = pd.read_csv('cs-test.csv', index_col=0)

sample_data = pd.read_csv('sampleEntry.csv', index_col=0)

EDA

data.head()
SeriousDlqin2yrsRevolvingUtilizationOfUnsecuredLinesageNumberOfTime30-59DaysPastDueNotWorseDebtRatioMonthlyIncomeNumberOfOpenCreditLinesAndLoansNumberOfTimes90DaysLateNumberRealEstateLoansOrLinesNumberOfTime60-89DaysPastDueNotWorseNumberOfDependents
110.7661274520.8029829120.0130602.0
200.9571514000.1218762600.040001.0
300.6581803810.0851133042.021000.0
400.2338103000.0360503300.050000.0
500.9072394910.02492663588.070100.0
test_data.head()
SeriousDlqin2yrsRevolvingUtilizationOfUnsecuredLinesageNumberOfTime30-59DaysPastDueNotWorseDebtRatioMonthlyIncomeNumberOfOpenCreditLinesAndLoansNumberOfTimes90DaysLateNumberRealEstateLoansOrLinesNumberOfTime60-89DaysPastDueNotWorseNumberOfDependents
1NaN0.8855194300.1775135700.040000.0
2NaN0.4632955700.5272379141.0150402.0
3NaN0.0432755900.6876485083.0120102.0
4NaN0.2803083810.9259613200.070200.0
5NaN1.0000002700.0199173865.040001.0
data.describe([0.01,0.1,0.25,.5,.75,.9,.99]).T
countmeanstdmin1%10%25%50%75%90%99%max
SeriousDlqin2yrs150000.00.0668400.2497460.00.00.0000000.0000000.0000000.0000000.0000001.0000001.0
RevolvingUtilizationOfUnsecuredLines150000.06.048438249.7553710.00.00.0029690.0298670.1541810.5590460.9812781.09295650708.0
age150000.052.29520714.7718660.024.033.00000041.00000052.00000063.00000072.00000087.000000109.0
NumberOfTime30-59DaysPastDueNotWorse150000.00.4210334.1927810.00.00.0000000.0000000.0000000.0000001.0000004.00000098.0
DebtRatio150000.0353.0050762037.8185230.00.00.0308740.1750740.3665080.8682541267.0000004979.040000329664.0
MonthlyIncome120269.06670.22123714384.6742150.00.02005.0000003400.0000005400.0000008249.00000011666.00000025000.0000003008750.0
NumberOfOpenCreditLinesAndLoans150000.08.4527605.1459510.00.03.0000005.0000008.00000011.00000015.00000024.00000058.0
NumberOfTimes90DaysLate150000.00.2659734.1693040.00.00.0000000.0000000.0000000.0000000.0000003.00000098.0
NumberRealEstateLoansOrLines150000.01.0182401.1297710.00.00.0000000.0000001.0000002.0000002.0000004.00000054.0
NumberOfTime60-89DaysPastDueNotWorse150000.00.2403874.1551790.00.00.0000000.0000000.0000000.0000000.0000002.00000098.0
NumberOfDependents146076.00.7572221.1150860.00.00.0000000.0000000.0000001.0000002.0000004.00000020.0
test_data.describe([0.01,0.1,0.25,.5,.75,.9,.99]).T
countmeanstdmin1%10%25%50%75%90%99%max
SeriousDlqin2yrs0.0NaNNaNNaNNaNNaNNaNNaNNaNNaNNaNNaN
RevolvingUtilizationOfUnsecuredLines101503.05.310000196.1560390.00.00.0030080.0301310.1525860.5642250.9833421.0886921821.0
age101503.052.40543614.77975621.024.033.00000041.00000052.00000063.00000072.00000087.00000104.0
NumberOfTime30-59DaysPastDueNotWorse101503.00.4537704.5384870.00.00.0000000.0000000.0000000.0000001.0000004.0000098.0
DebtRatio101503.0344.4750201632.5952310.00.00.0300580.1734230.3642600.8516191238.8000004963.00000268326.0
MonthlyIncome81400.06855.03559036508.6003750.00.02083.0000003408.0000005400.0000008200.00000011500.00000025916.010007727000.0
NumberOfOpenCreditLinesAndLoans101503.08.4535145.1441000.00.03.0000005.0000008.00000011.00000015.00000025.0000085.0
NumberOfTimes90DaysLate101503.00.2966914.5158590.00.00.0000000.0000000.0000000.0000000.0000003.0000098.0
NumberRealEstateLoansOrLines101503.01.0130741.1102530.00.00.0000000.0000001.0000002.0000002.0000004.0000037.0
NumberOfTime60-89DaysPastDueNotWorse101503.00.2703174.5035780.00.00.0000000.0000000.0000000.0000000.0000002.0000098.0
NumberOfDependents98877.00.7690461.1367780.00.00.0000000.0000000.0000001.0000002.0000004.0000043.0
data.shape

(150000, 11)

test_data.shape

(101503, 11)

data.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 150000 entries, 1 to 150000
Data columns (total 11 columns):
SeriousDlqin2yrs                        150000 non-null int64
RevolvingUtilizationOfUnsecuredLines    150000 non-null float64
age                                     150000 non-null int64
NumberOfTime30-59DaysPastDueNotWorse    150000 non-null int64
DebtRatio                               150000 non-null float64
MonthlyIncome                           120269 non-null float64
NumberOfOpenCreditLinesAndLoans         150000 non-null int64
NumberOfTimes90DaysLate                 150000 non-null int64
NumberRealEstateLoansOrLines            150000 non-null int64
NumberOfTime60-89DaysPastDueNotWorse    150000 non-null int64
NumberOfDependents                      146076 non-null float64
dtypes: float64(4), int64(7)
memory usage: 13.7 MB

test_data.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 101503 entries, 1 to 101503
Data columns (total 11 columns):
SeriousDlqin2yrs                        0 non-null float64
RevolvingUtilizationOfUnsecuredLines    101503 non-null float64
age                                     101503 non-null int64
NumberOfTime30-59DaysPastDueNotWorse    101503 non-null int64
DebtRatio                               101503 non-null float64
MonthlyIncome                           81400 non-null float64
NumberOfOpenCreditLinesAndLoans         101503 non-null int64
NumberOfTimes90DaysLate                 101503 non-null int64
NumberRealEstateLoansOrLines            101503 non-null int64
NumberOfTime60-89DaysPastDueNotWorse    101503 non-null int64
NumberOfDependents                      98877 non-null float64
dtypes: float64(5), int64(6)
memory usage: 9.3 MB

drop_duplicates

# 去重复
# 训练集去重复，测试集不要去重复
data.drop_duplicates(inplace=True)

data.info()

data.index = range(data.shape[0])

data.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 149391 entries, 1 to 150000
Data columns (total 11 columns):
SeriousDlqin2yrs                        149391 non-null int64
RevolvingUtilizationOfUnsecuredLines    149391 non-null float64
age                                     149391 non-null int64
NumberOfTime30-59DaysPastDueNotWorse    149391 non-null int64
DebtRatio                               149391 non-null float64
MonthlyIncome                           120170 non-null float64
NumberOfOpenCreditLinesAndLoans         149391 non-null int64
NumberOfTimes90DaysLate                 149391 non-null int64
NumberRealEstateLoansOrLines            149391 non-null int64
NumberOfTime60-89DaysPastDueNotWorse    149391 non-null int64
NumberOfDependents                      145563 non-null float64
dtypes: float64(4), int64(7)
memory usage: 13.7 MB
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 149391 entries, 0 to 149390
Data columns (total 11 columns):
SeriousDlqin2yrs                        149391 non-null int64
RevolvingUtilizationOfUnsecuredLines    149391 non-null float64
age                                     149391 non-null int64
NumberOfTime30-59DaysPastDueNotWorse    149391 non-null int64
DebtRatio                               149391 non-null float64
MonthlyIncome                           120170 non-null float64
NumberOfOpenCreditLinesAndLoans         149391 non-null int64
NumberOfTimes90DaysLate                 149391 non-null int64
NumberRealEstateLoansOrLines            149391 non-null int64
NumberOfTime60-89DaysPastDueNotWorse    149391 non-null int64
NumberOfDependents                      145563 non-null float64
dtypes: float64(4), int64(7)
memory usage: 12.5 MB

Checking Null values

data.isnull().mean()

SeriousDlqin2yrs                        0.000000
RevolvingUtilizationOfUnsecuredLines    0.000000
age                                     0.000000
NumberOfTime30-59DaysPastDueNotWorse    0.000000
DebtRatio                               0.000000
MonthlyIncome                           0.198207
NumberOfOpenCreditLinesAndLoans         0.000000
NumberOfTimes90DaysLate                 0.000000
NumberRealEstateLoansOrLines            0.000000
NumberOfTime60-89DaysPastDueNotWorse    0.000000
NumberOfDependents                      0.026160
dtype: float64

data.nunique()

SeriousDlqin2yrs                             2
RevolvingUtilizationOfUnsecuredLines    125728
age                                         86
NumberOfTime30-59DaysPastDueNotWorse        16
DebtRatio                               114194
MonthlyIncome                            13594
NumberOfOpenCreditLinesAndLoans             58
NumberOfTimes90DaysLate                     19
NumberRealEstateLoansOrLines                28
NumberOfTime60-89DaysPastDueNotWorse        13
NumberOfDependents                          13
dtype: int64

test_data.isnull().mean()

SeriousDlqin2yrs                        1.000000
RevolvingUtilizationOfUnsecuredLines    0.000000
age                                     0.000000
NumberOfTime30-59DaysPastDueNotWorse    0.000000
DebtRatio                               0.000000
MonthlyIncome                           0.198053
NumberOfOpenCreditLinesAndLoans         0.000000
NumberOfTimes90DaysLate                 0.000000
NumberRealEstateLoansOrLines            0.000000
NumberOfTime60-89DaysPastDueNotWorse    0.000000
NumberOfDependents                      0.025871
dtype: float64

# NumberOfDependents
data['NumberOfDependents'].fillna(int(data['NumberOfDependents'].mode()[0]), inplace=True)

data.isnull().mean()

SeriousDlqin2yrs                        0.000000
RevolvingUtilizationOfUnsecuredLines    0.000000
age                                     0.000000
NumberOfTime30-59DaysPastDueNotWorse    0.000000
DebtRatio                               0.000000
MonthlyIncome                           0.198207
NumberOfOpenCreditLinesAndLoans         0.000000
NumberOfTimes90DaysLate                 0.000000
NumberRealEstateLoansOrLines            0.000000
NumberOfTime60-89DaysPastDueNotWorse    0.000000
NumberOfDependents                      0.000000
dtype: float64

test_data['NumberOfDependents'].fillna(int(test_data['NumberOfDependents'].mode()[0]), inplace=True)

test_data.isnull().mean()

SeriousDlqin2yrs                        1.000000
RevolvingUtilizationOfUnsecuredLines    0.000000
age                                     0.000000
NumberOfTime30-59DaysPastDueNotWorse    0.000000
DebtRatio                               0.000000
MonthlyIncome                           0.198053
NumberOfOpenCreditLinesAndLoans         0.000000
NumberOfTimes90DaysLate                 0.000000
NumberRealEstateLoansOrLines            0.000000
NumberOfTime60-89DaysPastDueNotWorse    0.000000
NumberOfDependents                      0.000000
dtype: float64

# MonthlyIncome
data['MonthlyIncome'].fillna(int(data['MonthlyIncome'].mean()), inplace=True)

data.isnull().mean()

SeriousDlqin2yrs                        0.0
RevolvingUtilizationOfUnsecuredLines    0.0
age                                     0.0
NumberOfTime30-59DaysPastDueNotWorse    0.0
DebtRatio                               0.0
MonthlyIncome                           0.0
NumberOfOpenCreditLinesAndLoans         0.0
NumberOfTimes90DaysLate                 0.0
NumberRealEstateLoansOrLines            0.0
NumberOfTime60-89DaysPastDueNotWorse    0.0
NumberOfDependents                      0.0
dtype: float64

test_data['MonthlyIncome'].fillna(int(test_data['MonthlyIncome'].mean()), inplace=True)

test_data.isnull().mean()

SeriousDlqin2yrs                        1.0
RevolvingUtilizationOfUnsecuredLines    0.0
age                                     0.0
NumberOfTime30-59DaysPastDueNotWorse    0.0
DebtRatio                               0.0
MonthlyIncome                           0.0
NumberOfOpenCreditLinesAndLoans         0.0
NumberOfTimes90DaysLate                 0.0
NumberRealEstateLoansOrLines            0.0
NumberOfTime60-89DaysPastDueNotWorse    0.0
NumberOfDependents                      0.0
dtype: float64

Visualization

sns.countplot(x='SeriousDlqin2yrs',data=data)
plt.show()

在这里插入图片描述

f,ax = plt.subplots(figsize=(10, 10))
sns.heatmap(data.corr(), annot=True, linewidths=.5, fmt= '.3f',ax=ax)
plt.show()

在这里插入图片描述

auusual values

# ###################
# 训练集
data.loc[data['age']==0, 'age'] = int(data['age'].mean())


# 测试集也做同样的处理
test_data.loc[test_data['age']==0, 'age'] = int(test_data['age'].mean())


#划分训练集测试集
X = data.loc[:, data.columns!='SeriousDlqin2yrs'] 
y = data.loc[:,'SeriousDlqin2yrs']

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=116214)

# 上采样
smote = SMOTE(random_state=0)
s_X,s_y=smote.fit_sample(X_train,y_train)

LogisticRegression model

# 基础模型
LR = LogisticRegression()
LR.fit(X_train, y_train)
print('训练集精确度:{0}，测试集精确度:{1}\n'.format(LR.score(X_train, y_train), LR.score(X_test, y_test)))

print('predict 训练集auc:{0}，测试集auc:{1}\n'.format(roc_auc_score(s_y, LR.predict(s_X)),
                                                       roc_auc_score(y_test, LR.predict(X_test))))
print('predict_proba 训练集auc:{0}，测试集auc:{1}\n'.format(roc_auc_score(s_y, LR.predict_proba(s_X)[:,1]),
                                                               roc_auc_score(y_test, LR.predict_proba(X_test)[:,1])))

训练集精确度:0.933075，测试集精确度:0.9334666666666667

predict 训练集auc:0.5064304086881967，测试集auc:0.5078828573129228

predict_proba 训练集auc:0.6839255109393088，测试集auc:0.6871543091031289

# 基础模型2
LR_s = LogisticRegression()
LR_s.fit(s_X,s_y)
print('训练集精确度:{0}，测试集精确度:{1}\n'.format(LR_s.score(s_X,s_y), LR_s.score(X_test, y_test)))

print('predict 训练集auc:{0}，测试集auc:{1}\n'.format(roc_auc_score(s_y, LR_s.predict(s_X)),
                                                       roc_auc_score(y_test, LR_s.predict(X_test))))
print('predict_proba 训练集auc:{0}，测试集auc:{1}\n'.format(roc_auc_score(s_y, LR_s.predict_proba(s_X)[:,1]),
                                                               roc_auc_score(y_test, LR_s.predict_proba(X_test)[:,1])))
# FPR, TPR, _ = roc_curve(s_y, LR_s.predict_proba(s_X)[:,1])
# FPR_test, TPR_test, _ = roc_curve(y_test, LR_s.predict_proba(X_test)[:,1])
# print('训练集auc:{0}，测试集auc:{1}\n'.format(auc(FPR, TPR),auc(FPR_test, TPR_test)))

# print('训练集混淆矩阵:\n{0}，\n\n测试集混淆矩阵:\n{1}'.format(cm(s_y, y_pred_train_s,labels=[1,0]),
#                                                                cm(y_test, y_pred_s,labels=[1,0])))

skplt.metrics.plot_roc(y_test, pd.DataFrame(LR_s.predict_proba(X_test)), plot_micro=False, figsize=(6,6), plot_macro=False)

训练集精确度:0.6768451700485854，测试集精确度:0.8703333333333333

predict 训练集auc:0.6768451700485852，测试集auc:0.7351483859562692

predict_proba 训练集auc:0.7682496553604893，测试集auc:0.8034115289523501

在这里插入图片描述

# 随机网格搜索
LR_r = LogisticRegression()
param_dist = {
        'penalty':['l1', 'l2'],
        'C':[0, 0.01, 0.05, 0.1, 0.5, 1, 5, 10, 50, 100],
        'solver':['newton-cg', 'lbfgs', 'liblinear', 'sag', 'saga'],
        'max_iter':range(1,300,50),
        }
LR_random = RandomizedSearchCV(LR_r,param_distributions=param_dist,cv=5,n_jobs = -1)
LR_random.fit(s_X,s_y)

best_est_LR = LR_random.best_estimator_
print('最优参数:{0}\n'.format(best_est_LR))

print('训练集精确度:{0}，测试集精确度:{1}\n'.format(LR_random.score(s_X,s_y), LR_random.score(X_test, y_test)))

y_pred_train_sr = LR_random.predict(s_X)
y_pred_sr = LR_random.predict(X_test)
print('训练集auc:{0}，测试集auc:{1}\n'.format(roc_auc_score(s_y, y_pred_train_sr),roc_auc_score(y_test, y_pred_sr)))

print('训练集混淆矩阵:\n{0}，\n\n测试集混淆矩阵:\n{1}'.format(cm(s_y, y_pred_train_sr,labels=[1,0]),
                                                               cm(y_test, y_pred_sr,labels=[1,0])))

最优参数:LogisticRegression(C=50, class_weight=None, dual=False, fit_intercept=True,
                   intercept_scaling=1, l1_ratio=None, max_iter=251,
                   multi_class='auto', n_jobs=None, penalty='l2',
                   random_state=None, solver='lbfgs', tol=0.0001, verbose=0,
                   warm_start=False)

训练集精确度:0.6722322449271221，测试集精确度:0.722

训练集auc:0.672232244927122，测试集auc:0.6570985182928633

训练集混淆矩阵:
[[68407 43561]
 [29838 82130]]，

测试集混淆矩阵:
[[ 1161   833]
 [ 7507 20499]]

# 手动调参
auc_score = []
for i in ['newton-cg', 'lbfgs', 'liblinear', 'sag', 'saga']:
    LR_m = LogisticRegression(solver=i)
    LR_m.fit(s_X,s_y)
    auc_score.append(roc_auc_score(y_test, LR_m.predict_proba(X_test)[:, 1]))

print(max(auc_score))
plt.plot(['newton-cg', 'lbfgs', 'liblinear', 'sag', 'saga'],auc_score)
plt.show()

0.8034115289523501

在这里插入图片描述

# 手动调参
auc_score = []
for i in [0.1, 0.5, 1]:
    LR_m = LogisticRegression(solver='lbfgs', C=i)
    LR_m.fit(s_X,s_y)
    auc_score.append(roc_auc_score(y_test, LR_m.predict_proba(X_test)[:, 1]))
print(max(auc_score))
plt.plot([0.1, 0.5, 1],auc_score)
plt.show()

0.8072174013291749

在这里插入图片描述

# 手动调参
auc_score = []
for i in [0.01, 0.05, 0.1]:
    LR_m = LogisticRegression(solver='lbfgs', C=i)
    LR_m.fit(s_X,s_y)
    auc_score.append(roc_auc_score(y_test, LR_m.predict_proba(X_test)[:, 1]))
print(max(auc_score))
plt.plot([0.01, 0.05, 0.1],auc_score)
plt.show()

0.8072174013291749

在这里插入图片描述

# 手动调参
auc_score = []
for i in [50, 100, 150, 200]:
    LR_m = LogisticRegression(solver='lbfgs', C=0.1, max_iter=i)
    LR_m.fit(s_X,s_y)
    auc_score.append(roc_auc_score(y_test, LR_m.predict_proba(X_test)[:, 1]))
print(max(auc_score))
plt.plot([50, 100, 150, 200],auc_score)
plt.show()

0.8072174013291749

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(30, 120, 10):
    LR_m = LogisticRegression(solver='lbfgs', C=0.1, max_iter=i)
    LR_m.fit(s_X,s_y)
    auc_score.append(roc_auc_score(y_test, LR_m.predict_proba(X_test)[:, 1]))
print(list(range(30, 120, 10))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(30, 120, 10),auc_score)
plt.show()

80 0.807376433735972

在这里插入图片描述

best_est_LR = LogisticRegression(solver='lbfgs', C=0.1, max_iter=80).fit(s_X,s_y)

df_test = test_data.drop('SeriousDlqin2yrs', axis=1)
sample_data["Probability"] = best_est_LR.predict_proba(df_test)[:,1]
print(sample_data.head())
sample_data.to_csv("submission_LR.csv",index=False)

    Probability
Id             
1      0.403650
2      0.392559
3      0.419110
4      0.487821
5      0.472049

Random Forest model

# 基础模型
RFC = RandomForestClassifier()
RFC.fit(X_train, y_train)
print('训练集精确度:{0}，测试集精确度:{1}\n'.format(RFC.score(X_train, y_train), RFC.score(X_test, y_test)))

print('predict 训练集auc:{0}，测试集auc:{1}\n'.format(roc_auc_score(y_train, RFC.predict(X_train)),
                                                       roc_auc_score(y_test, RFC.predict(X_test))))
print('predict_proba 训练集auc:{0}，测试集auc:{1}\n'.format(roc_auc_score(y_train, RFC.predict_proba(X_train)[:,1]),
                                                            roc_auc_score(y_test, RFC.predict_proba(X_test)[:,1])))

训练集精确度:0.9995916666666667，测试集精确度:0.9359

predict 训练集auc:0.997585340973915，测试集auc:0.5895362836348795

predict_proba 训练集auc:0.9999771540267908，测试集auc:0.8437090354115978

# 基础模型2
RFC_s = RandomForestClassifier()
RFC_s.fit(s_X,s_y)
print('训练集精确度:{0}，测试集精确度:{1}\n'.format(RFC_s.score(s_X,s_y), RFC_s.score(X_test, y_test)))

print('predict 训练集auc:{0}，测试集auc:{1}\n'.format(roc_auc_score(s_y, RFC_s.predict(s_X)),
                                                       roc_auc_score(y_test, RFC_s.predict(X_test))))
print('predict_proba 训练集auc:{0}，测试集auc:{1}\n'.format(roc_auc_score(s_y, RFC_s.predict_proba(s_X)[:,1]),
                                                            roc_auc_score(y_test, RFC_s.predict_proba(X_test)[:,1])))

训练集精确度:0.9993390968848243，测试集精确度:0.8908333333333334

predict 训练集auc:0.9993390968848243，测试集auc:0.6839441591216555

predict_proba 训练集auc:0.9999081457017259，测试集auc:0.8237175910363382

# 随机网格搜索
RFC_r = RandomForestClassifier()
param_dist = {
        'n_estimators':[10, 20, 30, 40, 50, 100, 150, 200, 300],
        'max_depth':range(2, 10, 1),
        'min_samples_leaf':range(2, 10, 1),
        }
RFC_random = RandomizedSearchCV(RFC_r,param_distributions=param_dist,cv=5,n_jobs = -1)
RFC_random.fit(s_X,s_y)

best_est_RFC = RFC_random.best_estimator_
print('最优参数:{0}\n'.format(best_est_RFC))

print('训练集精确度:{0}，测试集精确度:{1}\n'.format(RFC_random.score(s_X,s_y), RFC_random.score(X_test, y_test)))

y_pred_train_sr = RFC_random.predict(s_X)
y_pred_sr = RFC_random.predict(X_test)
print('训练集auc:{0}，测试集auc:{1}\n'.format(auc(s_y, y_pred_train_sr),auc(y_test, y_pred_sr)))

print('训练集混淆矩阵:\n{0}，\n\n测试集混淆矩阵:\n{1}'.format(cm(s_y, y_pred_train_sr,labels=[1,0]),
                                                               cm(y_test, y_pred_sr,labels=[1,0])))

最优参数:RandomForestClassifier(bootstrap=True, ccp_alpha=0.0, class_weight=None,
                       criterion='gini', max_depth=9, max_features='auto',
                       max_leaf_nodes=None, max_samples=None,
                       min_impurity_decrease=0.0, min_impurity_split=None,
                       min_samples_leaf=3, min_samples_split=2,
                       min_weight_fraction_leaf=0.0, n_estimators=10,
                       n_jobs=None, oob_score=False, random_state=None,
                       verbose=0, warm_start=False)

训练集精确度:0.8192925526193122，测试集精确度:0.7796780347401184

训练集auc:0.8192925526193122，测试集auc:0.7544857947353516

训练集混淆矩阵:
[[94898 16658]
 [23660 87896]]，

测试集混淆矩阵:
[[ 1489   564]
 [ 6019 21807]]

# 手动调参
auc_score = []
for i in range(50, 350, 50):
    RF_m = RandomForestClassifier(n_estimators=i,max_depth=5,min_samples_split=200,min_samples_leaf=100,
                                  max_features='sqrt',bootstrap=0.7,n_jobs=-1,random_state=10,)
    RF_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, RF_m.predict_proba(X_test)[:, 1]))
print(list(range(50, 350, 50))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(50, 350, 50),auc_score)
plt.show()

250 0.8640378842017733

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(50, 350, 50):
    RF_m = RandomForestClassifier(n_estimators=i,max_depth=5,min_samples_split=200,min_samples_leaf=100,
                                  max_features='sqrt',bootstrap=0.7,n_jobs=-1,random_state=10,)
    RF_m.fit(s_X, s_y)
    auc_score.append(roc_auc_score(y_test, RF_m.predict_proba(X_test)[:, 1]))
print(list(range(50, 350, 50))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(50, 350, 50),auc_score)
plt.show()

250 0.8389906830396209

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(200, 300, 10):
    RF_m = RandomForestClassifier(n_estimators=i,max_depth=5,min_samples_split=200,min_samples_leaf=100,
                                  max_features='sqrt',bootstrap=0.7,n_jobs=-1,random_state=10,)
    RF_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, RF_m.predict_proba(X_test)[:, 1]))
print(list(range(200, 300, 10))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(200, 300, 10),auc_score)
plt.show()

230 0.8640656490645973

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(2,10,1):
    RF_m = RandomForestClassifier(n_estimators=230,max_depth=i,min_samples_split=200,min_samples_leaf=100,
                                  max_features='sqrt',bootstrap=0.7,n_jobs=-1,random_state=10,)
    RF_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, RF_m.predict_proba(X_test)[:, 1]))
print(list(range(2,10,1))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(2,10,1),auc_score)
plt.show()

9 0.8677676427124693

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(9,15,1):
    RF_m = RandomForestClassifier(n_estimators=230,max_depth=i,min_samples_split=200,min_samples_leaf=100,
                                  max_features='sqrt',bootstrap=0.7,n_jobs=-1,random_state=10,)
    RF_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, RF_m.predict_proba(X_test)[:, 1]))
print(list(range(9,15,1))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(9,15,1),auc_score)
plt.show()

12 0.8684788028299709

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(50,350,50):
    RF_m = RandomForestClassifier(n_estimators=230,max_depth=12,min_samples_split=i,min_samples_leaf=100,
                                  max_features='sqrt',bootstrap=0.7,n_jobs=-1,random_state=10,)
    RF_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, RF_m.predict_proba(X_test)[:, 1]))
print(list(range(50,350,50))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(50,350,50),auc_score)
plt.show()

50 0.8684788028299709

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(10,55,10):
    RF_m = RandomForestClassifier(n_estimators=230,max_depth=12,min_samples_split=i,min_samples_leaf=100,
                                  max_features='sqrt',bootstrap=0.7,n_jobs=-1,random_state=10,)
    RF_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, RF_m.predict_proba(X_test)[:, 1]))
print(list(range(10,55,10))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(10,55,10),auc_score)
plt.show()

10 0.8684788028299709

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(2,11,1):
    RF_m = RandomForestClassifier(n_estimators=230,max_depth=12,min_samples_split=i,min_samples_leaf=100,
                                  max_features='sqrt',bootstrap=0.7,n_jobs=-1,random_state=10,)
    RF_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, RF_m.predict_proba(X_test)[:, 1]))
print(list(range(2,11,1))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(2,11,1),auc_score)
plt.show()

2 0.8684788028299709

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(10, 110, 10):
    RF_m = RandomForestClassifier(n_estimators=230,max_depth=12,min_samples_split=2,min_samples_leaf=i,
                                  max_features='sqrt',bootstrap=0.7,n_jobs=-1,random_state=10,)
    RF_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, RF_m.predict_proba(X_test)[:, 1]))
print(list(range(10, 110, 10))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(10, 110, 10),auc_score)
plt.show()

40 0.8687270122873082

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(36, 46, 2):
    RF_m = RandomForestClassifier(n_estimators=230,max_depth=12,min_samples_split=2,min_samples_leaf=i,
                                  max_features='sqrt',bootstrap=0.7,n_jobs=-1,random_state=10,)
    RF_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, RF_m.predict_proba(X_test)[:, 1]))
print(list(range(36, 46, 2))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(36, 46, 2),auc_score)
plt.show()

36 0.8691479440821931

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(30, 50, 2):
    RF_m = RandomForestClassifier(n_estimators=230,max_depth=12,min_samples_split=2,min_samples_leaf=i,
                                  max_features='sqrt',bootstrap=0.7,n_jobs=-1,random_state=10,)
    RF_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, RF_m.predict_proba(X_test)[:, 1]))
print(list(range(30, 50, 2))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(30, 50, 2),auc_score)
plt.show()

36 0.8691479440821931

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(1,11,1):
    RF_m = RandomForestClassifier(n_estimators=230,max_depth=12,min_samples_split=2,min_samples_leaf=36,
                                  max_features=i/10,bootstrap=0.7,n_jobs=-1,random_state=10,)
    RF_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, RF_m.predict_proba(X_test)[:, 1]))
print(list(range(1,11,1))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(1,11,1),auc_score)
plt.show()

3 0.8691479440821931

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(1,11,1):
    RF_m = RandomForestClassifier(n_estimators=230,max_depth=12,min_samples_split=2,min_samples_leaf=36,
                                  max_features=0.3,bootstrap=i,n_jobs=-1,random_state=10,)
    RF_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, RF_m.predict_proba(X_test)[:, 1]))
print(list(range(1,11,1))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(1,11,1),auc_score)
plt.show()

1 0.8691479440821931

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(1,15,2):
    RF_m = RandomForestClassifier(n_estimators=230,max_depth=12,min_samples_split=2,min_samples_leaf=36,
                                  max_features=0.3,bootstrap=0.1,n_jobs=-1,random_state=10,class_weight={1:i})
    RF_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, RF_m.predict_proba(X_test)[:, 1]))
print(list(range(1,15,2))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(1,15,2),auc_score)
plt.show()

1 0.8691479440821931

在这里插入图片描述

best_est_RFC = RandomForestClassifier(n_estimators=230,max_depth=12,min_samples_split=2,min_samples_leaf=36,
                              max_features=0.3,bootstrap=0.1,n_jobs=-1,random_state=10,)
best_est_RFC.fit(X_train, y_train)
print(roc_auc_score(y_test, best_est_RFC.predict_proba(X_test)[:, 1]))

0.8691479440821931

df_test = test_data.drop('SeriousDlqin2yrs', axis=1)
sample_data["Probability"] = best_est_RFC.predict_proba(df_test)[:,1]
print(sample_data.head())
sample_data.to_csv("submission_RFC.csv",index=True)

    Probability
Id             
1      0.079361
2      0.036505
3      0.016210
4      0.064601
5      0.126260

Grandient Boosting model

# 基础模型
GB = GradientBoostingClassifier()
GB.fit(X_train, y_train)
print('训练集精确度:{0}，测试集精确度:{1}\n'.format(GB.score(X_train, y_train), GB.score(X_test, y_test)))

print('predict 训练集auc:{0}，测试集auc:{1}\n'.format(roc_auc_score(y_train, GB.predict(X_train)),
                                                       roc_auc_score(y_test, GB.predict(X_test))))
print('predict_proba 训练集auc:{0}，测试集auc:{1}\n'.format(roc_auc_score(y_train, GB.predict_proba(X_train)[:,1]),
                                                            roc_auc_score(y_test, GB.predict_proba(X_test)[:,1])))

训练集精确度:0.9389916666666667，测试集精确度:0.9377333333333333

predict 训练集auc:0.6017569654221069，测试集auc:0.5916827107760473

predict_proba 训练集auc:0.8685441072547121，测试集auc:0.8677365023729332

# 基础模型2
GB_s = GradientBoostingClassifier()
GB_s.fit(s_X,s_y)
print('训练集精确度:{0}，测试集精确度:{1}\n'.format(GB_s.score(s_X,s_y), GB_s.score(X_test, y_test)))

print('predict 训练集auc:{0}，测试集auc:{1}\n'.format(roc_auc_score(s_y, GB_s.predict(s_X)),
                                                       roc_auc_score(y_test, GB_s.predict(X_test))))
print('predict_proba 训练集auc:{0}，测试集auc:{1}\n'.format(roc_auc_score(s_y, GB_s.predict_proba(s_X)[:,1]),
                                                            roc_auc_score(y_test, GB_s.predict_proba(X_test)[:,1])))

训练集精确度:0.8654704915690197，测试集精确度:0.8668

predict 训练集auc:0.8654704915690197，测试集auc:0.7248715725122952

predict_proba 训练集auc:0.9424107237355832，测试集auc:0.8360759633753794

# 手动调参
auc_score = []
for i in range(50, 350, 50):
    GB_m = GradientBoostingClassifier(learning_rate=0.1,n_estimators=i,subsample=0.7,min_samples_split=200,
                                        min_samples_leaf=100,max_depth=3,random_state=10,max_features='sqrt')
    GB_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, GB_m.predict_proba(X_test)[:, 1]))
print(list(range(50, 350, 50))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(50, 350, 50),auc_score)
plt.show()

250 0.8684184382039928

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(200, 300, 10):
    GB_m = GradientBoostingClassifier(learning_rate=0.1,n_estimators=i,subsample=0.7,min_samples_split=200,
                                        min_samples_leaf=100,max_depth=3,random_state=10,max_features='sqrt')
    GB_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, GB_m.predict_proba(X_test)[:, 1]))
print(list(range(200, 300, 10))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(200, 300, 10),auc_score)
plt.show()

250 0.8684184382039928

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(2,10,1):
    GB_m = GradientBoostingClassifier(learning_rate=0.1,n_estimators=250,subsample=0.7,min_samples_split=200,
                                        min_samples_leaf=100,max_depth=i,random_state=10,max_features='sqrt')
    GB_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, GB_m.predict_proba(X_test)[:, 1]))
print(list(range(2,10,1))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(2,10,1),auc_score)
plt.show()

4 0.8694069872976782

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(1,11,1):
    GB_m = GradientBoostingClassifier(learning_rate=0.1,n_estimators=250,subsample=0.7,min_samples_split=200,
                                        min_samples_leaf=100,max_depth=4,random_state=10,max_features=i/10)
    GB_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, GB_m.predict_proba(X_test)[:, 1]))
print(list(range(1,11,1))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(1,11,1),auc_score)
plt.show()

3 0.8694069872976782

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(10, 310, 50):
    GB_m = GradientBoostingClassifier(learning_rate=0.1,n_estimators=250,subsample=0.7,min_samples_split=200,
                                        min_samples_leaf=i,max_depth=4,random_state=10,max_features=0.3)
    GB_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, GB_m.predict_proba(X_test)[:, 1]))
print(list(range(10, 310, 50))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(10, 310, 50),auc_score)
plt.show()

110 0.8694922785209158

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(70, 130, 10):
    GB_m = GradientBoostingClassifier(learning_rate=0.1,n_estimators=250,subsample=0.7,min_samples_split=200,
                                        min_samples_leaf=i,max_depth=4,random_state=10,max_features=0.3)
    GB_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, GB_m.predict_proba(X_test)[:, 1]))
print(list(range(70, 130, 10))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(70, 130, 10),auc_score)
plt.show()

90 0.8697006000505265

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(10, 300, 50):
    GB_m = GradientBoostingClassifier(learning_rate=0.1,n_estimators=250,subsample=0.7,min_samples_split=i,
                                        min_samples_leaf=90,max_depth=4,random_state=10,max_features=0.3)
    GB_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, GB_m.predict_proba(X_test)[:, 1]))
print(list(range(10, 300, 50))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(10, 300, 50),auc_score)
plt.show()

260 0.869784512431818

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(250, 410, 50):
    GB_m = GradientBoostingClassifier(learning_rate=0.1,n_estimators=250,subsample=0.7,min_samples_split=i,
                                        min_samples_leaf=90,max_depth=4,random_state=10,max_features=0.3)
    GB_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, GB_m.predict_proba(X_test)[:, 1]))
print(list(range(250, 410, 50))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(250, 410, 50),auc_score)
plt.show()

250 0.8696629236420251

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(250, 300, 10):
    GB_m = GradientBoostingClassifier(learning_rate=0.1,n_estimators=250,subsample=0.7,min_samples_split=i,
                                        min_samples_leaf=90,max_depth=4,random_state=10,max_features=0.3)
    GB_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, GB_m.predict_proba(X_test)[:, 1]))
print(list(range(250, 300, 10))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(250, 300, 10),auc_score)
plt.show()

260 0.869784512431818

在这里插入图片描述

# 手动调参
auc_score = []
for i in range(1,11,1):
    GB_m = GradientBoostingClassifier(learning_rate=0.1,n_estimators=250,subsample=i/10,min_samples_split=260,
                                        min_samples_leaf=90,max_depth=4,random_state=10,max_features=0.3)
    GB_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, GB_m.predict_proba(X_test)[:, 1]))
print(list(range(1,11,1))[auc_score.index(max(auc_score))], max(auc_score))
plt.plot(range(1,11,1),auc_score)
plt.show()

7 0.869784512431818

在这里插入图片描述

# 手动调参
auc_score = []
for i in [10, 42, 100, 116214]:
    GB_m = GradientBoostingClassifier(learning_rate=0.1,n_estimators=250,subsample=0.7,min_samples_split=260,
                                        min_samples_leaf=90,max_depth=4,random_state=i,max_features=0.3)
    GB_m.fit(X_train, y_train)
    auc_score.append(roc_auc_score(y_test, GB_m.predict_proba(X_test)[:, 1]))
print([10, 42, 100, 116214][auc_score.index(max(auc_score))], max(auc_score))
plt.plot([10, 42, 100, 116214],auc_score)
plt.show()

10 0.869784512431818

在这里插入图片描述

# 基础模型
GB_m = GradientBoostingClassifier(learning_rate=0.1,n_estimators=250,subsample=0.7,min_samples_split=260,
                                        min_samples_leaf=90,max_depth=4,random_state=10,max_features=0.3)
GB_m.fit(X_train, y_train)
print('训练集精确度:{0}，测试集精确度:{1}\n'.format(GB_m.score(X_train, y_train), GB_m.score(X_test, y_test)))

print('predict 训练集auc:{0}，测试集auc:{1}\n'.format(roc_auc_score(y_train, GB_m.predict(X_train)),
                                                       roc_auc_score(y_test, GB_m.predict(X_test))))
print('predict_proba 训练集auc:{0}，测试集auc:{1}\n'.format(roc_auc_score(y_train, GB_m.predict_proba(X_train)[:,1]),
                                                            roc_auc_score(y_test, GB_m.predict_proba(X_test)[:,1])))

训练集精确度:0.9400916666666667，测试集精确度:0.9371666666666667

predict 训练集auc:0.6089339590765261，测试集auc:0.5925436990826797

predict_proba 训练集auc:0.874079427703056，测试集auc:0.869784512431818

# 基础模型
GB_m = GradientBoostingClassifier(learning_rate=0.05,n_estimators=500,subsample=0.7,min_samples_split=260,
                                        min_samples_leaf=90,max_depth=4,random_state=10,max_features=0.3)
GB_m.fit(X_train, y_train)
print('训练集精确度:{0}，测试集精确度:{1}\n'.format(GB_m.score(X_train, y_train), GB_m.score(X_test, y_test)))

print('predict 训练集auc:{0}，测试集auc:{1}\n'.format(roc_auc_score(y_train, GB_m.predict(X_train)),
                                                       roc_auc_score(y_test, GB_m.predict(X_test))))
print('predict_proba 训练集auc:{0}，测试集auc:{1}\n'.format(roc_auc_score(y_train, GB_m.predict_proba(X_train)[:,1]),
                                                            roc_auc_score(y_test, GB_m.predict_proba(X_test)[:,1])))

训练集精确度:0.940275，测试集精确度:0.9373666666666667

predict 训练集auc:0.6067785694888352，测试集auc:0.5903218295893178

predict_proba 训练集auc:0.8745757049324849，测试集auc:0.8691238322551743

best_est_GB = GradientBoostingClassifier(learning_rate=0.1,n_estimators=250,subsample=0.7,min_samples_split=260,
                                        min_samples_leaf=90,max_depth=4,random_state=10,max_features=0.3)
best_est_GB.fit(X_train, y_train)

GradientBoostingClassifier(ccp_alpha=0.0, criterion='friedman_mse', init=None,
                           learning_rate=0.1, loss='deviance', max_depth=4,
                           max_features=0.3, max_leaf_nodes=None,
                           min_impurity_decrease=0.0, min_impurity_split=None,
                           min_samples_leaf=90, min_samples_split=260,
                           min_weight_fraction_leaf=0.0, n_estimators=250,
                           n_iter_no_change=None, presort='deprecated',
                           random_state=10, subsample=0.7, tol=0.0001,
                           validation_fraction=0.1, verbose=0,
                           warm_start=False)

df_test = test_data.drop('SeriousDlqin2yrs', axis=1)
sample_data["Probability"] = best_est_GB.predict_proba(df_test)[:,1]
print(sample_data.head())
sample_data.to_csv("submission_GB.csv",index=True)

    Probability
Id             
1      0.072539
2      0.043227
3      0.013998
4      0.064475
5      0.083119
小步积
关注
  • 0
    点赞
  • 踩
  • 13
    收藏
    觉得还不错? 一键收藏
  • 2
    评论
专栏目录
《大数据机器学习实践探索》---- 使用pyspark 进行kaggle比赛Give me some credit数据集的建模与分析（3. 随机森林进行二分类）
shiter编写程序的艺术
04-19 540
文章大纲随机森林 原理参考文献 随机森林 原理 原理 　　随机森林是由许多决策树构成，是一种有监督机器学习方法，可以用于分类和回归，通过合并汇总来自个体决策树的结果来进行预测，采用多数选票作为分类结果，采用预测结果平均值作为回归结果。 “森林”的概念很好理解，“随机”是针对森林中的每一颗决策树，有两种含义： 第一种随机是数据采样随机，构建决策树的训练数据集通过有放回的随机采样，并且只会选择一定百分比的样本，这样可以在数据集合存在噪声点、异常点的情况下，有些决策树的构造过程中不会选择到这些噪声点、异常点从而
来自Kaggle的Give Me Some Credit数据
11-12
数据来自Kaggle的Give Me Some Credit，有15万条的样本数据，下图可以看到这份数据的大致情况。 数据属于个人消费类贷款，只考虑信用评分最终实施时能够使用到的数据应从如下一些方面获取数据： – 基本属性：包括了借款人当时的年龄。 – 偿债能力：包括了借款人的月收入、负债比率。 – 信用往来：两年内35-59天逾期次数、两年内60-89天逾期次数、两年内90 天或高于90天逾期的次数。 – 财产状况：包括了开放式信贷和贷款数量、不动产贷款或额度数量。 – 贷款属性：暂无。 – 其他因素：包括了借款人的家属数量（不包括本人在内）。 – 时间窗口：自变量的观察窗口为过去两年，因变量表现窗口为未来两年。
kaggle上的Give Me Some Credit数据用于数据挖掘分析
07-30
kaggle上的Give Me Some Credit数据的挖掘分析，结合信用评分卡的建立原理，从数据的预处理、变量选择、建模分析到创建信用评分，创建了一个简单的信用评分系统
Kaggle:Give me some credit（1）
weixin_43822286的博客
04-24 2204
背景 信用评分算法，对默认可能性进行猜测，这是银行用来判断贷款是否应该被授予的方法，完成一个评分卡，通过预测某人在未来两年将会经历财务危机的可能性来提高信用评分的效果，帮助贷款人做出最好的决策 本项目主要为申请者评分模型的开发过程。 数据集介绍 数据来源：数据来自Kaggle，cs-training.csv是有15万条的样本数据，下图可以看到这份数据的大致情况。下载地址为：https://www....
超详细用Python进行信用评分卡建模【kaggle的give me some credit数据集】【风控建模】
热门推荐
weixin_43570097的博客
05-22 1万+
1 信用评分卡模型简介 信用评分卡是一个通过个人数据对其还款能力、还款意愿进行定量评估的系统。在消费金融行业，信用评分卡主要有三种（A、B、C卡）： A卡：申请评分卡，贷前阶段使用，在客户获取期，建立信用风险评分，预测客户带来违约风险的概率大小； B卡：行为评分卡，贷中阶段引入，在客户借款处理期，建立申请风险评分模型，预测客户违约拖欠的风险概率，我们的B卡采用的是T+1离线计算出来的，针对复借用户特别有效，针对首借用户，B卡自动转为申请评分；B卡另外的价值还在于用户授信到借款之间的时间比较长的话，通过B卡能
机器学习kaggle案例：风控评分卡模型（Give_Me_Some_Credit）
weixin_42662126的博客
08-21 8995
kaggle链接:https://www.kaggle.com/c/GiveMeSomeCredit ipynb文件：https://github.com/824024445/KaggleCases/blob/master/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0kaggle%E6%A1%88%E4%BE%8B%EF%BC%9A%E9%A3%8E%E6%8E%A7%...
Kaggle give me some credit数据集
05-21
Give Me Some Credit是Kaggle上关于信用评分的项目，通过改进信用评分技术，预测未来两年借款人会遇到财务困境的可能性。银行在市场经济中发挥关键作用。 他们决定谁可以获得融资，以及以何种条件进行投资决策。 为了市场和社会的运作，个人和公司需要获得信贷。信用评分算法可以猜测违约概率，这是银行用于确定是否应授予贷款的方法。目标是建立一个借款人可以用来帮助做出最佳财务决策的模型。
该信用卡评分案例，建模数据来自Kaggle竞赛网的GiveMeSomeCredit数据集。.zip
10-23
Kaggle是一个全球知名的数据科学竞赛平台，它提供了各种数据集供数据科学家进行研究和建模，其中包括GiveMeSomeCredit数据集。这个数据集主要用于信用卡申请的信用评分，通过分析个人的信用历史、财务状况等多维度...
kaggle的give me some credit数据集,包含training文件和data dictionary
07-21
"Give Me Some Credit"是Kaggle上一个著名的数据科学竞赛，主要目的是预测个人信贷违约的可能性。这个数据集提供了一种实践机器学习算法，特别是信用评分模型的机会。在这个任务中，我们将深入探讨其中的关键知识点...
kaggle “give me some credit”数据集
02-22
kaggle “give me some credit”数据集
Give Me Some Credit
07-30
这是本人从kaggle官网下载下来的Give Me Some Credit竞赛的相关数据，自己目前也在学习这一块内容，希望大家一起学习。
kaggle 上“give me some credit” 信用比赛的原始数据
04-28
这是从kaggle 上下载的“give me some credit” 信用比赛的原始数据，用于个人信用评估相关机器学习模型
Give Me Some Credit kaggle 源数据
04-16
Give Me Some Credit kaggle 源数据，供下载研究学习使用，请勿商业用途
kaggle/give_me_some_credit数据中的测试数据
04-15
搬运自kaggle平台比赛中的上give_me_some_credit数据中的测试数据。
基于lightgbm的kaggle比赛实践：Give me some credit
谷雨的博客
10-16 3337
0 背景介绍 Give Me Some Credit https://www.kaggle.com/c/GiveMeSomeCredit/overview，是Kaggle上关于信用评分的项目，通过改进信用评分技术，预测未来两年借款人会遇到财务困境的可能性。并以此为依据来决定是否给予借贷人信用授权。目标是建立帮助银行做出最佳财务借贷决策的模型。今天这 数据类型如下： 其中：SeriousDlqin2yrs代表过去两年内的情况，也是test集要预测的字段。 第一部分：导入需要的包和数据 impor.
金融风控信用评分卡建模（Kaggle give me credit数据集）
最新发布
weixin_43488159的博客
04-23 2101
风控信用评分卡
Kaggle|Give Me Some Credit信用卡评分建模分析(R语言)
Iceforest的博客
04-26 1万+
1.目的    本文是基于Kaggle|Give Me Some Credit项目(数据地址:https://www.kaggle.com/c/GiveMeSomeCredit),通过对消费者的人口特征、信用历史记录、交易记录等大量数据进行系统的分析、挖掘数据蕴含的行为模式、信用特征，发展出预测行的模式，结合信用卡评分的构建原理，采用R语言完成数据的清洗，主要包括缺失数据的填充、异常的删除和数据的...
Kaggle数据科学竞赛-数据分析流程笔记（以Give Me Some Credit为例）（未完待续~）
in_void的博客
04-08 3363
以Kaggle上的Give Me Some Credit赛题为例，详细阐述数据分析的基本方法与思路。
Kaggle:Give me some credit（2）
weixin_43822286的博客
05-05 1084
3.特征工程 2.1 数据分箱 利用caret包中的createDataPartition（数据分割功能）函数将数据随机分成相同的两份。训练集和测集的结果是平衡的，大致都为6.6%左右，因此可以采用这份切割的数据进行建模及预测。 set.seed(123) splitIndex<-createDataPartition(traindata$y,time=1,p=0.5,list=FALS...
解释DATA_HUB['kaggle_house_train'] = ( #@save DATA_URL + 'kaggle_house_pred_train.csv',
05-01
HUB的字典，并在其中添加一个键值对，键为'kag这段代码是在Python中创建一个名为DATA_HUB的字典，并在其中添加一个键值对，键为'kaggle这段代码是在Python中创建一个名为DATA_HUB的字典，并在其中添加一个键值对，键...
评论 2
添加红包

请填写红包祝福语或标题

个

红包个数最小为10个

元

红包金额最低5元

当前余额3.43元 前往充值 >
需支付：10.00元
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付元
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明：

1.余额是钱包充值的虚拟货币，按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载，可以购买VIP、付费专栏及课程。

余额充值