信用评分卡模型案例（R语言实现）

相似文件 换一批

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

1. library(xlsx)
   
2. dat<-read.xlsx2("credit.xlsx",1,colClasses = NA)
   
3. summary(dat)
   
4. dat[,1:6]<-sapply(dat[,1:6],function(x) {x[x==999]<-NA;return(x)} )
   
5. dat<-dat[,-11]
   
6. 
   
7. library(smbinning)
   
8. library(prettyR)
   
9. 
   
10. dat1<-dat
    
11. dat1[,4]<-dat1[,4]-dat1[,3]
    
12. table(dat1[,4])
    
13. dat1[,4]<-factor(dat1[,4],levels=c(1,2),labels=c("其他","已婚"))
    
14. colnames(dat1)<-c("age","income","child","marital","dur_live",
    
15.                   "dur_work","housetype","nation","cardtype","loan")
    
16. summary(dat1)
    
17. write.csv(dat1,"dat1.csv",row.names = F)

复制代码

credit.xlsx文件： credit.xlsx (156.4 KB)

2018-10-19 16:51:04 上传

1. dat1<-read.csv("dat1.csv")
   
2. 
   
3. ##盖帽法函数
   
4. block<-function(x,lower=T,upper=T){
   
5.     if(lower){
   
6.         q1<-quantile(x,0.01)
   
7.         x[x<=q1]<-q1
   
8.     }
   
9.     if(upper){
   
10.         q99<-quantile(x,0.99)
    
11.         x[x>q99]<-q99
    
12.     }
    
13.     return(x)
    
14. }
    
15. 
    
16. dat1$loan<-as.numeric(!as.logical(dat1$loan))
    
17. 
    
18. ##age
    
19. boxplot(age~loan,data=dat1,horizontal=T, frame=F,
    
20.         col="lightgray",main="Distribution")
    
21. age<-smbinning(dat1,"loan","age")
    
22. age$ivtable
    
23. par(mfrow=c(2,2))
    
24. smbinning.plot(age,option="dist",sub="年龄")
    
25. smbinning.plot(age,option="WoE",sub="年龄")
    
26. smbinning.plot(age,option="goodrate",sub="年龄")
    
27. smbinning.plot(age,option="badrate",sub="年龄")
    
28. par(mfrow=c(1,1))
    
29. age$iv
    
30. cred_iv<-c("年龄"=age$iv)
    
31. 
    
32. ##income
    
33. boxplot(income~loan,data=dat1,horizontal=T, frame=F,
    
34.         col="lightgray",main="Distribution")
    
35. dat1$income<-block(dat1$income)
    
36. boxplot(income~loan,data=dat1,horizontal=T, frame=F,
    
37.         col="lightgray",main="Distribution")
    
38. income<-smbinning(dat1,"loan","income")
    
39. income$ivtable
    
40. smbinning.plot(income,option="WoE",sub="收入")
    
41. income$iv
    
42. cred_iv<-c(cred_iv,"收入"=income$iv)
    
43. 
    
44. ##child
    
45. boxplot(child~loan,data=dat1,horizontal=T, frame=F,
    
46.         col="lightgray",main="Distribution")
    
47. dat1$child<-block(dat1$child)
    
48. child<-smbinning(dat1,"loan","child")
    
49. child$ivtable
    
50. smbinning.plot(child,option="WoE",sub="孩子数量")
    
51. child$iv
    
52. cred_iv<-c(cred_iv,"孩子数量"=child$iv)
    
53. 
    
54. ##marital
    
55. xtab(~marital+loan,data=dat1,chisq=T)
    
56. marital<-smbinning.factor(dat1,"loan","marital")
    
57. marital$ivtable
    
58. smbinning.plot(marital,option="WoE",sub="婚姻状态")
    
59. marital$iv
    
60. cred_iv<-c(cred_iv,"婚姻状态"=marital$iv)
    
61. 
    
62. ##dur_live
    
63. boxplot(dur_live~loan,data=dat1,horizontal=T,
    
64.         frame=F, col="lightgray",main="Distribution")
    
65. t.test(dur_live~loan,data=dat1)
    
66. dur_live<-smbinning(dat1,"loan","dur_live")
    
67. dur_live
    
68. 
    
69. ##dur_work
    
70. boxplot(dur_work~loan,data=dat1,horizontal=T,
    
71.         frame=F, col="lightgray",main="Distribution")
    
72. t.test(dur_work~loan,data=dat1)
    
73. dur_work<-smbinning(dat1,"loan","dur_work")
    
74. dur_work$ivtable
    
75. smbinning.plot(dur_work,option="WoE",sub="在现工作时间")
    
76. dur_work$iv
    
77. cred_iv<-c(cred_iv,"在现工作时间"=dur_work$iv)
    
78. 
    
79. ##housetype
    
80. xtab(~housetype+loan,data=dat1,chisq=T)
    
81. housetype<-smbinning.factor(dat1,"loan","housetype")
    
82. housetype$ivtable
    
83. smbinning.plot(housetype,option="WoE",sub="住房类型")
    
84. housetype$iv
    
85. cred_iv<-c(cred_iv,"住房种类"=housetype$iv)
    
86. 
    
87. ##nation
    
88. xtab(~nation+loan,data=dat1,chisq=T)
    
89. nation<-smbinning.factor(dat1,"loan","nation")
    
90. nation$ivtable
    
91. smbinning.plot(nation,option="WoE",sub="国籍")
    
92. nation$iv
    
93. cred_iv<-c(cred_iv,"国籍"=nation$iv)
    
94. 
    
95. ##cardtype
    
96. xtab(~cardtype+loan,data=dat1,chisq=T)
    
97. cardtype<-smbinning.factor(dat1,"loan","cardtype")
    
98. cardtype$ivtable
    
99. smbinning.plot(cardtype,option="WoE",sub="信用卡类型")
    
100. cardtype$iv
     
101. cred_iv<-c(cred_iv,"信用卡类型"=cardtype$iv)
     
102. 
     
103. barplot(cred_iv,main="各变量信息值")
     
104. 
     
105. dat2<-dat1
     
106. dat2<-smbinning.gen(dat2,age,"glage")
     
107. dat2<-smbinning.gen(dat2,income,"glincome")
     
108. dat2<-smbinning.gen(dat2,child,"glchild")
     
109. dat2<-smbinning.factor.gen(dat2,marital,"glmarital")
     
110. dat2<-smbinning.gen(dat2,dur_work,"gldur_work")
     
111. dat2<-smbinning.factor.gen(dat2,housetype,"glhousetype")
     
112. dat2<-smbinning.factor.gen(dat2,nation,"glnation")
     
113. dat2<-smbinning.factor.gen(dat2,cardtype,"glcardtype")
     
114. 
     
115. dat3<-dat2[,c(11:18,10)]
     
116. 
     
117. write.csv(dat3,"dat3.csv",row.names = F)

复制代码

dat1.zip (21.42 KB) 本附件包括：

• dat1.csv

2018-10-19 16:52:42 上传

1. dat3<-read.csv("dat3.csv")
   
2. 
   
3. cred_mod<-glm(loan~. ,data=dat3,family=binomial())
   
4. summary(cred_mod)
   
5. 
   
6. cre_scal<-smbinning.scaling(cred_mod,pdo=45,score=800,odds=50)
   
7. cre_scal$logitscaled
   
8. cre_scal$minmaxscore
   
9. 
   
10. 
    
11. dat4<-smbinning.scoring.gen(smbscaled=cre_scal, dataset=dat3)
    
12. boxplot(Score~loan,data=dat4,horizontal=T, frame=F,
    
13.         col="lightgray",main="Distribution")
    
14. 
    
15. scaledcard<-cre_scal$logitscaled[[1]][-1,c(1,2,6)]
    
16. scaledcard[,1]<-c(rep("年龄",5),rep("收入",3),
    
17.                   rep("孩子数量",2),rep("婚否",2),rep("在现工作时间",5),
    
18.                   rep("住房类型",3),rep("国籍",8),rep("信用卡类型",7))
    
19. scaledcard
    
20. write.csv(scaledcard,"card.csv",row.names = F)

复制代码

知识点3.zip (12.73 KB) 本附件包括：

• card.csv
• creditcard.R
• dat3.csv

2018-10-19 16:53:52 上传

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏3 回帖

关键词：信用评分卡 评分卡模型 信用评分 评分卡 R语言 评分卡 信用风险评分卡 信用评分卡

东方祥 发表于 2018-10-19 16:55
credit.xlsx文件：

thanks LZ for your kindness

thanks a lot for sharing the code, but could not call library(xlsx), got this error message

Error: package or namespace load failed for ‘xlsx’:
.onLoad failed in loadNamespace() for 'rJava', details:
  call: fun(libname, pkgname)
  error: No CurrentVersion entry in Software/JavaSoft registry! Try re-installing Java and make sure R and Java have matching architectures.

any idea??

hifinecon 发表于 2018-10-19 23:26
thanks a lot for sharing the code, but could not call library(xlsx), got this error message

Error ...

装个java的jdk还是什么的就好了，没有java环境

家庭人口数—孩子数量  为什么等于2就是已婚，等于1就是其他呢？  不应该孩子数量不为0 就是已婚吗  孩子都有了啊，还有等于1也有可能是已婚丧偶啊。这样处理感觉不是太好啊...

哎呦不错哦~ 发表于 2018-10-20 20:24
装个java的jdk还是什么的就好了，没有java环境

tried it and it still did not work, same error. I'll just use other way to read the xlsx file. thanks for your help.

哎呦不错哦~ 发表于 2018-10-20 20:27
家庭人口数—孩子数量  为什么等于2就是已婚，等于1就是其他呢？  不应该孩子数量不为0 就是已婚吗  孩子都 ...

yes, I also have the doubt, but this is just an example. we could modify it accordingly. cheers!

hifinecon 发表于 2018-10-20 23:12
yes, I also have the doubt, but this is just an example. we could modify it accordingly. cheers!

可是孩子数量的值的类是小于10个啊，不能进行smbinning分箱操作啊。