【数据挖掘案例实践】Loan Status

一、LoanStatus案例介绍

• 数据集变量：
  

• 预测变量：
  Loan_Status贷款是否成功

• 评价指标：
  Cross_validation Score (ROC曲线–AUC得分)

• 数据集大小
  训练集样本数：614
  测试集样本数：367

二、查看数据集

1.导入Python数据挖掘库

import pandas as pdimport numpy as npimport matplotlib.pyplot as plt

2.查看数据集

df.head()

df.describe()

三、变量统计分析

1.查看Property_Area变量

 temp1 = df['Property_Area'].value_counts() print(temp1)

temp1.plot(kind='bar')

2.查看ApplicantIncome字段（连续变量）

# 绘制ApplicantIncome的直方图df['ApplicantIncome'].hist(bins=30)

#查看ApplicantIncome 箱线图df.boxplot(column='ApplicantIncome')

#查看不同教育水平的收入情况df.boxplot(column='ApplicantIncome',by='Education')

#查看不同教育水平及性别的收入情况df.boxplot(column='ApplicantIncome',by=['Education','Gender'])

3. 查看LoanAmount字段（连续变量）

#绘制LoanAmount直方图df['LoanAmount'].hist(bins=40)

#绘制箱线图df.boxplot(column='LoanAmount')

#查看不同教育水平的人的借款金额箱线图df.boxplot(column='LoanAmount',by='Education')

#看不同性别的人的借款情况df.boxplot(column='LoanAmount',by='Gender')

#不同教育水平&不同性别df.boxplot(column='LoanAmount',by=['Gender','Education'])

4.基于信用情况获得贷款的可能性（Credit_History）

temp1 = df['Credit_History'].value_counts(ascending=True)

temp2=df.pivot_table(values='Loan_Status',index=['Credit_History'],    aggfunc=lambda x:x.map({'Y':1,'N':0}).mean())

import matplotlib.pyplot as pltfig=plt.figure(figsize=(8,4))ax1=fig.add_subplot(121)ax1.set_xlabel('Credit_History')ax1.set_ylabel('Count of Applicants')ax1.set_title("Applicants by Credit_History")temp1.plot(kind='bar')ax2=fig.add_subplot(122)ax2.set_xlabel('Credit_History')ax2.set_ylabel('Probability of getting loan')ax2.set_title("Probability of getting loan by Credit_History")temp2.plot(kind='bar')

temp3=pd.crosstab(df['Credit_History'],df['Loan_Status'])temp3.plot(kind='bar',stacked=True,color=['red','blue'],grid=False)

temp4=pd.crosstab([df['Credit_History'],df['Gender']],df['Loan_Status'])temp4.plot(kind='bar',stacked=True,color=['red','blue'],grid=False)

由图可知，历史信用会对贷款结果产生较大的影响

5. 基于婚姻情况获得贷款的可能性

temp4=df['Married'].value_counts(ascending=True)

temp5=df.pivot_table(values='Loan_Status',index=['Married'],    aggfunc=lambda x:x.map({'Y':1,'N':0}).mean())

四、数据预处理

• 在一些变量中存在缺失值，基于丢失的值的数量估计这些缺失值，并评估变量的重要性
• applicantIncome和LoanAmount存在极值，应适当处理
• 处理非数值领域Gender，Property_Area，Married，Education，Dependents

1.查看所有变量中的缺失值个数

df.apply(lambda x:sum(x.isnull()),axis=0)

2.填补Self_Employed的缺失值

df['Self_Employed'].value_counts()df['Self_Employed'].fillna('No',inplace=True)

3.处理Credit_History的缺失值

df['Credit_History'].value_counts()df['Credit_History'].fillna(1,inplace=True)

处理后：

4.处理Loan_Amount_Term的缺失值

df['Loan_Amount_Term'].describe()df['Loan_Amount_Term'].fillna(df['Loan_Amount_Term'].notnull().mean(),inplace=True)

5.创建数据透视表，提供我们所有Education和Self_employed变量的唯一值分组的中位数

table=df.pivot_table(values='LoanAmount',index='Self_Employed',columns='Education',aggfunc=np.median)def fage(x):    return table.loc[x['Self_Employed'],x['Education']]#Replace missing valuesdf['LoanAmount'].fillna(df[df['LoanAmount'].isnull()].apply(fage,axis=1),inplace=True)df['LoanAmount'].hist(bins=40)

6.applicantIncome 将申请者和共同申请者两者的收入结合起来作为总收入，并采取相同的对数变换

df['TotalIncome']=df['ApplicantIncome']+df['CoapplicantIncome']df['TotalIncome_log']=np.log(df['TotalIncome'])df['TotalIncome_log'].hist(bins=20)

7.哑变量处理
1）处理Gender缺失值

df['Gender'].fillna('Male',inplace=True)

五、建模分析