Lending Club——构建贷款违约预测模型

1.本项目需解决的问题

    本项目通过利用P2P平台Lending Club的贷款数据,进行机器学习,构建贷款违约预测模型,对新增贷款申请人进行预测是否会违约,从而决定是否放款。


2.建模思路

    以下为本次项目的工作流程。


3.场景解析

    贷款申请人向Lending Club平台申请贷款时,Lending Club平台通过线上或线下让客户填写贷款申请表,收集客户的基本信息,这里包括申请人的年龄、性别、婚姻状况、学历、贷款金额、申请人财产情况等信息,通常来说还会借助第三方平台如征信机构或FICO等机构的信息。通过这些信息属性来做线性回归 ,生成预测模型,Lending Club平台可以通过预测判断贷款申请是否会违约,从而决定是否向申请人发放贷款。

1)首先,我们的场景是通过用户的历史行为(如历史数据的多维特征和贷款状态是否违约)来训练模型,通过这个模型对新增的贷款人“是否具有偿还能力,是否具有偿债意愿”进行分析,预测贷款申请人是否会发生违约贷款。这是一个监督学习的场景,因为已知了特征以及贷款状态是否违约(目标列),我们判定贷款申请人是否违约是一个二元分类问题,可以通过一个分类算法来处理,这里选用逻辑斯蒂回归(Logistic Regression)。

2)观察数据集发现部分数据是半结构化数据,需要进行特征抽象。

    现对该业务场景进行总结如下:

  •     根据历史记录数据学习并对贷款是否违约进行预测,监督学习场景,选择逻辑斯蒂回归(Logistic Regression)算法
  • 数据为半结构化数据,需要进行特征抽象

4.数据预处理(Pre-Processing Data)

     本次项目数据集来源于Lending Club Statistics,具体为2018年第一季Lending Club平台发生借贷的业务数据

 数据预览


查看每列属性缺失值的比例

check_null = data.isnull().sum().sort_values(ascending=False)/float(len(data)) 
print(check_null[check_null > 0.2]) # 查看缺失比例大于20%的属性。

    从上面信息可以发现,本次数据集缺失值较多的属性对我们模型预测意义不大,例如id和member_id以及url等。因此,我们直接删除这些没有意义且缺失值较多的属性。此外,如果缺失值对属性来说是有意义的,还得细分缺失值对应的属性是数值型变量或是分类类型变

thresh_count = len(data)*0.4 # 设定阀值
data = data.dropna(thresh=thresh_count, axis=1) #若某一列数据缺失的数量超过阀值就会被删除

再将处理后的数据转化为csv

data.to_csv('loans_2018q1_ml.csv', index = False)
loans = pd.read_csv('loans_2018q1_ml.csv') 
loans.dtypes.value_counts() # 分类统计数据类型

loans.shape

(107866, 103)

同值化处理   

    如果一个变量大部分的观测都是相同的特征,那么这个特征或者输入变量就是无法用来区分目标时间。

loans = loans.loc[:,loans.apply(pd.Series.nunique) != 1]
loans.shape

(107866, 96)

缺失值处理——分类变量   

objectColumns = loans.select_dtypes(include=["object"]).columns
loans[objectColumns].isnull().sum().sort_values(ascending=False)
loans[objectColumns]

loans['int_rate'] = loans['int_rate'].str.rstrip('%').astype('float')
loans['revol_util'] = loans['revol_util'].str.rstrip('%').astype('float')
objectColumns = loans.select_dtypes(include=["object"]).columns  

    我们可以调用missingno库来快速评估数据缺失的情况。

msno.matrix(loans[objectColumns])  # 缺失值可视化

    从图中可以直观看出变量“last_pymnt_d”、“emp_title”、“emp_length”缺失值较多。

    这里我们先用‘unknown’来填充。

objectColumns = loans.select_dtypes(include=["object"]).columns 
loans[objectColumns] = loans[objectColumns].fillna("Unknown") 

缺失值处理——数值变量

numColumns = loans.select_dtypes(include=[np.number]).columns

pd.set_option('display.max_columns', len(numColumns))
loans[numColumns].tail()
loans.drop([107864, 107865], inplace =True)
    这里使用可sklearn的Preprocessing模块,参数strategy选用most_frequent,采用众数插补的方法填充缺失值。
imr = Imputer(missing_values='NaN', strategy='most_frequent', axis=0)  #  axis=0  针对列来处理
imr = imr.fit(loans[numColumns])
loans[numColumns] = imr.transform(loans[numColumns])

    这样缺失值就已经处理完。


数据过滤

print(objectColumns)

    将以上重复或对构建预测模型没有意义的属性进行删

drop_list = ['sub_grade', 'emp_title', 'issue_d', 'title', 'zip_code', 'addr_state', 'earliest_cr_line',
       'initial_list_status', 'last_pymnt_d', 'next_pymnt_d', 'last_credit_pull_d', 'disbursement_method']

loans.drop(drop_list, axis=1, inplace=True)
loans.select_dtypes(include = ['object']).shape

(107866, 8)

5.特征工程(Feature Engineering)

特征衍生

    Lending Club平台中,"installment"代表贷款每月分期的金额,我们将'annual_inc'除以12个月获得贷款申请人的月收入金额,然后再把"installment"(月负债)与('annual_inc'/12)(月收入)相除生成新的特征'installment_feat',新特征'installment_feat'代表客户每月还款支出占月收入的比,'installment_feat'的值越大,意味着贷款人的偿债压力越大,违约的可能性越大。
loans['installment_feat'] = loans['installment'] / ((loans['annual_inc']+1) / 12)

特征抽象(Feature Abstraction)

def coding(col, codeDict):

    colCoded = pd.Series(col, copy=True)
    for key, value in codeDict.items():
        colCoded.replace(key, value, inplace=True)

    return colCoded

#把贷款状态LoanStatus编码为违约=1, 正常=0:

loans["loan_status"] = coding(loans["loan_status"], {'Current':0,'Issued':0,'Fully Paid':0,'In Grace Period':1,'Late (31-120 days)':1,'Late (16-30 days)':1,'Charged Off':1})

print( '\nAfter Coding:')

pd.value_counts(loans["loan_status"])

        贷款状态可视化


loans.select_dtypes(include=["object"]).head()


    首先,我们对变量“emp_length”、"grade"进行特征抽象化。

# 有序特征的映射
mapping_dict = {
    "emp_length": {
        "10+ years": 10,
        "9 years": 9,
        "8 years": 8,
        "7 years": 7,
        "6 years": 6,
        "5 years": 5,
        "4 years": 4,
        "3 years": 3,
        "2 years": 2,
        "1 year": 1,
        "< 1 year": 0,
        "Unknown": 0
    },
    "grade":{
        "A": 1,
        "B": 2,
        "C": 3,
        "D": 4,
        "E": 5,
        "F": 6,
        "G": 7
    }
}

loans = loans.replace(mapping_dict) 
loans[['emp_length','grade']].head() 

    再对剩余特征进行One-hot编码。

n_columns = ["home_ownership", "verification_status", "application_type","purpose", "term"] 
dummy_df = pd.get_dummies(loans[n_columns]) # 用get_dummies进行one hot编码
loans = pd.concat([loans, dummy_df], axis=1) #当axis = 1的时候,concat就是行对齐,然后将不同列名称的两张表合并

    再清除掉原来的属性。

loans = loans.drop(n_columns, axis=1)
loans.info()

    这样,就已经将所有类型为object的变量作了转化。

col = loans.select_dtypes(include=['int64','float64']).columns
col = col.drop('loan_status') #剔除目标变量

loans_ml_df = loans  # 复制数据至变量loans_ml_df

特征缩放(Feature Scaling

     我们采用的是标准化的方法,调用scikit-learn模块preprocessing的子模块StandardScaler。
sc =StandardScaler()  # 初始化缩放器
loans_ml_df[col] =sc.fit_transform(loans_ml_df[col])  #对数据进行标准化

特征选择(Feature Selecting

    目的:首先,优先选择与目标相关性较高的特征;其次,去除不相关特征可以降低学习的难度
#构建X特征变量和Y目标变量
x_feature = list(loans_ml_df.columns)
x_feature.remove('loan_status')
x_val = loans_ml_df[x_feature]
y_val = loans_ml_df['loan_status']
len(x_feature) # 查看初始特征集合的数量
103
     首先,选出与目标变量相关性较高的特征。这里采用的是Wrapper方法,通过暴力的递归特征消除 (Recursive Feature Elimination)方法筛选30个与目标变量相关性最强的特征,逐步剔除特征从而达到首次降维,自变量从103个降到30个。
# 建立逻辑回归分类器
model = LogisticRegression()
# 建立递归特征消除筛选器
rfe = RFE(model, 30) #通过递归选择特征,选择30个特征
rfe = rfe.fit(x_val, y_val)
# 打印筛选结果
print(rfe.n_features_)
print(rfe.estimator_ )
print(rfe.support_)
print(rfe.ranking_) #ranking 为 1代表被选中,其他则未被代表未被选中
col_filter = x_val.columns[rfe.support_] #通过布尔值筛选首次降维后的变量
col_filter

Filter

    在第一次降维的基础上,通过皮尔森相关性图谱找出冗余特征并将其剔除;同时,可以通过相关性图谱进一步引导我们选择特征的方向。

colormap = plt.cm.viridis
plt.figure(figsize=(12,12))
plt.title('Pearson Correlation of Features', y=1.05, size=15)
sns.heatmap(loans_ml_df[col_filter].corr(),linewidths=0.1,vmax=1.0, square=True, cmap=colormap, linecolor='white', annot=True)

drop_col = ['funded_amnt', 'funded_amnt_inv', 'out_prncp', 'out_prncp_inv', 'total_pymnt_inv', 'total_rec_prncp',
           'num_actv_rev_tl', 'num_rev_tl_bal_gt_0', 'home_ownership_RENT', 'application_type_Joint App',
            'term_ 60 months', 'purpose_debt_consolidation', 'verification_status_Source Verified', 'home_ownership_OWN', 
             'verification_status_Verified',]
col_new = col_filter.drop(drop_col) #剔除冗余特征

len(col_new)  # 特征子集包含的变量从30个降维至15个。

15

Embedded
    下面需要对特征的权重有一个正确的评判和排序,可以通过特征重要性排序来挖掘哪些变量是比较重要的,降低学习难度,最终达到优化模型计算的目的。这里,我们采用的是随机森林算法判定特征的重要性,工程实现方式采用scikit-learn的featureimportances 的方法。
names = loans_ml_df[col_new].columns
clf=RandomForestClassifier(n_estimators=10,random_state=123) #构建分类随机森林分类器
clf.fit(x_val[col_new], y_val) #对自变量和因变量进行拟合
for feature in zip(names, clf.feature_importances_):
    print(feature)

plt.style.use('ggplot')

## feature importances 可视化##
importances = clf.feature_importances_
feat_names = names
indices = np.argsort(importances)[::-1]
fig = plt.figure(figsize=(20,6))
plt.title("Feature importances by RandomTreeClassifier")
plt.bar(range(len(indices)), importances[indices], color='lightblue',  align="center")
plt.step(range(len(indices)), np.cumsum(importances[indices]), where='mid', label='Cumulative')
plt.xticks(range(len(indices)), feat_names[indices], rotation='vertical',fontsize=14)
plt.xlim([-1, len(indices)])
plt.show()

# 下图是根据特征在特征子集中的相对重要性绘制的排序图,这些特征经过特征缩放后,其特征重要性的和为1.0。
# 由下图我们可以得出的结论:基于决策树的计算,特征子集上最具判别效果的特征是“total_pymnt”。

6.模型训练

处理样本不均衡

    前面已提到,目标变量“loans_status”正常和违约两种类别的数量差别较大,会对模型学习造成困扰。我们采用过采样的方法来处理样本不均衡问题,具体操作使用的是SMOTE(Synthetic Minority Oversampling Technique),SMOET的基本原理是:采样最邻近算法,计算出每个少数类样本的K个近邻,从K个近邻中随机挑选N个样本进行随机线性插值,构造新的少数样本,同时将新样本与原数据合成,产生新的训练集。

# 构建自变量和因变量
X = loans_ml_df[col_new]
y = loans_ml_df["loan_status"]

n_sample = y.shape[0]
n_pos_sample = y[y == 0].shape[0]
n_neg_sample = y[y == 1].shape[0]
print('样本个数:{}; 正样本占{:.2%}; 负样本占{:.2%}'.format(n_sample,
                                                   n_pos_sample / n_sample,
                                                   n_neg_sample / n_sample))
print('特征维数:', X.shape[1])
# 处理不平衡数据
sm = SMOTE(random_state=42)    # 处理过采样的方法
X, y = sm.fit_sample(X, y)
print('通过SMOTE方法平衡正负样本后')
n_sample = y.shape[0]
n_pos_sample = y[y == 0].shape[0]
n_neg_sample = y[y == 1].shape[0]
print('样本个数:{}; 正样本占{:.2%}; 负样本占{:.2%}'.format(n_sample,
                                                   n_pos_sample / n_sample,
                                                   n_neg_sample / n_sample))

构建分类器训练

    本次项目我们采用交叉验证法划分数据集,将数据划分为3部分:训练集(training set)、验证集(validation set)和测试集(test set)。让模型在训练集进行学习,在验证集上进行参数调优,最后使用测试集数据评估模型的性能。

    模型调优我们采用网格搜索调优参数(grid search),通过构建参数候选集合,然后网格搜索会穷举各种参数组合,根据设定评定的评分机制找到最好的那一组设置。

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.3, random_state = 0) # random_state = 0 每次切分的数据都一样
# 构建参数组合
param_grid = {'C': [0.01,0.1, 1, 10, 100, 1000,],
                            'penalty': [ 'l1', 'l2']}
# C:Inverse of regularization strength; must be a positive float. Like in support vector machines, smaller values specify stronger regularization.

grid_search = GridSearchCV(LogisticRegression(),  param_grid, cv=10) # 确定模型LogisticRegression,和参数组合param_grid ,cv指定10折
grid_search.fit(X_train, y_train) # 使用训练集学习算法

print("Best parameters: {}".format(grid_search.best_params_))
print("Best cross-validation score: {:.5f}".format(grid_search.best_score_))

print("Best estimator:\n{}".format(grid_search.best_estimator_)) # grid_search.best_estimator_ 返回模型以及他的所有参数(包含最优参数)

现在使用经过训练和调优后的模型在测试集上测试。

y_pred = grid_search.predict(X_test)
print("Test set accuracy score: {:.5f}".format(accuracy_score(y_test, y_pred,)))

Test set accuracy score: 0.66064

print(classification_report(y_test, y_pred))

roc_auc = roc_auc_score(y_test, y_pred)
print("Area under the ROC curve : %f" % roc_auc)

Area under the ROC curve : 0.660654


总结

    最后结果不太理想,实际工作中还要做特征分箱处理,计算IV值和WOE编码也是需要的。模型评估方面也有不足,这为以后的工作提供了些经验








  • 18
    点赞
  • 205
    收藏
    觉得还不错? 一键收藏
  • 10
    评论
lending club 贷款数据 2018年第二季度的贷款数据 "id","member_id","loan_amnt","funded_amnt","funded_amnt_inv","term","int_rate","installment","grade","sub_grade","emp_title","emp_length","home_ownership","annual_inc","verification_status","issue_d","loan_status","pymnt_plan","url","desc","purpose","title","zip_code","addr_state","dti","delinq_2yrs","earliest_cr_line","inq_last_6mths","mths_since_last_delinq","mths_since_last_record","open_acc","pub_rec","revol_bal","revol_util","total_acc","initial_list_status","out_prncp","out_prncp_inv","total_pymnt","total_pymnt_inv","total_rec_prncp","total_rec_int","total_rec_late_fee","recoveries","collection_recovery_fee","last_pymnt_d","last_pymnt_amnt","next_pymnt_d","last_credit_pull_d","collections_12_mths_ex_med","mths_since_last_major_derog","policy_code","application_type","annual_inc_joint","dti_joint","verification_status_joint","acc_now_delinq","tot_coll_amt","tot_cur_bal","open_acc_6m","open_act_il","open_il_12m","open_il_24m","mths_since_rcnt_il","total_bal_il","il_util","open_rv_12m","open_rv_24m","max_bal_bc","all_util","total_rev_hi_lim","inq_fi","total_cu_tl","inq_last_12m","acc_open_past_24mths","avg_cur_bal","bc_open_to_buy","bc_util","chargeoff_within_12_mths","delinq_amnt","mo_sin_old_il_acct","mo_sin_old_rev_tl_op","mo_sin_rcnt_rev_tl_op","mo_sin_rcnt_tl","mort_acc","mths_since_recent_bc","mths_since_recent_bc_dlq","mths_since_recent_inq","mths_since_recent_revol_delinq","num_accts_ever_120_pd","num_actv_bc_tl","num_actv_rev_tl","num_bc_sats","num_bc_tl","num_il_tl","num_op_rev_tl","num_rev_accts","num_rev_tl_bal_gt_0","num_sats","num_tl_120dpd_2m","num_tl_30dpd","num_tl_90g_dpd_24m","num_tl_op_past_12m","pct_tl_nvr_dlq","percent_bc_gt_75","pub_rec_bankruptcies","tax_liens","tot_hi_cred_lim","total_bal_ex_mort","total_bc_limit","total_il_high_credit_limit","revol_bal_joint","sec_app_earliest_cr_line","sec_app_inq_last_6mths","sec_app_mort_acc","sec_app_open_acc","sec_app_revol_util","sec_app_open_act_il","sec_app_num_rev
评论 10
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
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
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