Scorecard R

  library(openxlsx)
  library(smbinning)
  library(prettyR)
  # dat<-read.xlsx("/home/star/balack_namelist/scorecard/CreditScoreModel-master/cs-training.xlsx",1)
  dat<-read.xlsx("/home/star/balack_namelist/scorecard/CreditScoreModel-master/credit.xlsx",1)
  
  summary(dat)
  head(dat,2)
  str(dat)
  
  #去掉最后一列 
  dat<-dat[,-11]
  #dat<-dat[,-which(names(dat)%in%c("X1"))]#删除第一列方法2
  
  #生成新列——婚姻状况列
  dat1<-dat
  dat1[,4]<-dat1[,4]-dat1[,3]
  table(dat1[,4])
  dat1[,4]<-factor(dat1[,4],levels=c(1,2),labels=c("其他","已婚"))
  #重命名列
  colnames(dat1)<-c("age","income","child","marital","dur_live",
                    "dur_work","housetype","nation","cardtype","loan")
  summary(dat1)
  dat1[,c(4,7,8,9)]<-lapply(dat1[,c(4,7,8,9)],as.factor)
  ?lapply
  
  ##盖帽法函数
  block<-function(x,lower=T,upper=T){
    if(lower){
      q1<-quantile(x,0.01)
      x[x<=q1]<-q1
    }
    if(upper){
      q99<-quantile(x,0.99)
      x[x>q99]<-q99
    }
    return(x)
  }
  
  
  anyNA(dat1)
  #y中0和1互换(yuce weiyuegailv)
  dat1$loan<-as.numeric(!as.logical(dat1$loan))
  #dat1$loan <- 1- dat1$loan
  
  ##对每个变量分箱，求IV  
  ##age 
  boxplot(age~loan,data=dat1,horizontal=T, frame=F, 
          col="lightgray",main="Distribution")
  age<-smbinning(dat1,"loan","age")
  
  write.csv(age$ivtable,"/home/star/balack_namelist/scorecard/CreditScoreModel-master/ivtable.csv",row.names = F)
  
  age$ivtable
  par(mfrow=c(2,2))
  smbinning.plot(age,option="dist",sub="年龄")
  smbinning.plot(age,option="WoE",sub="年龄")
  smbinning.plot(age,option="goodrate",sub="年龄")
  smbinning.plot(age,option="badrate",sub="年龄")
  par(mfrow=c(1,1))
  age$iv
  cred_iv<-c("年龄"=age$iv)
  
  ##income
  boxplot(income~loan,data=dat1,horizontal=T, frame=F, 
          col="lightgray",main="Distribution")
  dat1$income<-block(dat1$income)
  boxplot(income~loan,data=dat1,horizontal=T, frame=F, 
          col="lightgray",main="Distribution")
  income<-smbinning(dat1,"loan","income")
  income$ivtable
  smbinning.plot(income,option="WoE",sub="收入")
  income$iv
  cred_iv<-c(cred_iv,"收入"=income$iv)
  
  ##child
  boxplot(child~loan,data=dat1,horizontal=T, frame=F, 
          col="lightgray",main="Distribution")
  dat1$child<-block(dat1$child)
  child<-smbinning(dat1,"loan","child")
  child$ivtable
  smbinning.plot(child,option="WoE",sub="孩子数量")
  child$iv
  cred_iv<-c(cred_iv,"孩子数量"=child$iv)
  
  ##marital
  xtab(~marital+loan,data=dat1,chisq=T)   
  marital<-smbinning.factor(dat1,"loan","marital")
  marital$ivtable
  smbinning.plot(marital,option="WoE",sub="婚姻状况")
  marital$iv
  cred_iv<-c(cred_iv,"婚姻状况"=marital$iv)
  
  ##dur_live
  boxplot(dur_live~loan,data=dat1,horizontal=T, 
          frame=F, col="lightgray",main="Distribution")
  t.test(dur_live~loan,data=dat1)
  dur_live<-smbinning(dat1,"loan","dur_live")
  dur_live
  
  ##dur_work
  boxplot(dur_work~loan,data=dat1,horizontal=T, 
          frame=F, col="lightgray",main="Distribution")
  t.test(dur_work~loan,data=dat1)
  dur_work<-smbinning(dat1,"loan","dur_work")
  dur_work$ivtable
  smbinning.plot(dur_work,option="WoE",sub="在现工作时间")
  dur_work$
    cred_iv<-c(cred_iv,"在现工作时间"=dur_work$iv)
  
  ##housetype
  xtab(~housetype+loan,data=dat1,chisq=T)
  #对分类变量最优分值
  housetype<-smbinning.factor(dat1,"loan","housetype")
  housetype$ivtable
  smbinning.plot(housetype,option="WoE",sub="住房类型")
  housetype$iv
  cred_iv<-c(cred_iv,"住房种类"=housetype$iv)
  
  ##nation
  xtab(~nation+loan,data=dat1,chisq=T)
  nation<-smbinning.factor(dat1,"loan","nation")
  nation$ivtable
  smbinning.plot(nation,option="WoE",sub="国籍")
  nation$iv
  cred_iv<-c(cred_iv,"国籍"=nation$iv)
  
  
  ##cardtype
  xtab(~cardtype+loan,data=dat1,chisq=T)
  cardtype<-smbinning.factor(dat1,"loan","cardtype")
  cardtype$ivtable
  smbinning.plot(cardtype,option="WoE",sub="信用卡类型")
  cardtype$iv
  cred_iv<-c(cred_iv,"信用卡类型"=cardtype$iv)
  
  
  barplot(cred_iv,main="各变量信息量")
  #生成分箱后的新列
  dat2<-dat1
  dat2<-smbinning.gen(dat2,age,"glage")
  dat2<-smbinning.gen(dat2,income,"glincome")
  dat2<-smbinning.gen(dat2,child,"glchild")
  dat2<-smbinning.factor.gen(dat2,marital,"glmarital")
  dat2<-smbinning.gen(dat2,dur_work,"gldur_work")
  dat2<-smbinning.factor.gen(dat2,housetype,"glhousetype")
  dat2<-smbinning.factor.gen(dat2,nation,"glnation")
  dat2<-smbinning.factor.gen(dat2,cardtype,"glcardtype")
  summary(dat2)
  dat3<-dat2[,c(11:18,10)]
  dat3
  cred_mod<-glm(loan~. ,data=dat3,family=binomial())
  summary(cred_mod)
  #打分
  ?smbinning.scaling
  cre_scal<-smbinning.scaling(cred_mod,pdo=45,score=800,odds=50)
  cre_scal
  cre_scal$minmaxscore
  cre_scal$logitscaled
  
  #生成每行对应的分值
  dat4<-smbinning.scoring.gen(smbscaled=cre_scal, dataset=dat3)
  par(mar=c(5,5,5,5)) # the plot is too large
  boxplot(Score~loan,data=dat4,horizontal=T, frame=F, 
          col="lightgray",main="Distribution")
  
  scaledcard<-cre_scal$logitscaled[[1]][-1,c(1,2,6)]
  scaledcard[,1]<-c(rep("年龄",5),rep("收入",3),
                    rep("孩子数量",2),rep("婚否",2),rep("在现工作时间",5),
                    rep("住房类型",3),rep("国籍",8),rep("信用卡类型",7))
  scaledcard
  write.csv(dat4,"/home/star/balack_namelist/scorecard/CreditScoreModel-master/card.csv",row.names = F)