Credit-Risk-Model/credit_riskmodel_germandata.R at master · aslanbm/Credit-Risk-Model · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146

dat=read.csv("file:///E:/chrome downloads/german_credit_data (1).csv")
head(dat)            #watch first 6 row of the data
dat = dat[,-1]
#know more about the data
#summary of the data
summary(dat)

#check the rate of missing values
miss = function(x){sum(is.na(x))/length(x)*100}
apply(dat,2,miss)

#pay attention to the unbalanced y-value
plot(dat[,10])

#devide them by y-value
good = dat[which(dat[,10]=="good"),]
bad = dat[which(dat[,10]=="bad"),]

####data exploring####
# 2.1 age & housing
#age
#win.graph(width = 8,height = 2.5)
par(mfrow=c(1,3))
hist(good[,1],xlab="age",ylab="",main="",ylim=c(0,160))
hist(bad[,1],xlab="age",ylab="",main="",ylim=c(0,160))
hist(dat[,1],xlab="age",ylab="",main="")

#housing
plot(good[,4],xlab="housing",ylab="",main="",ylim=c(0,500))
plot(bad[,4],xlab="housing",ylab="",main="",ylim=c(0,500))
plot(dat[,4],xlab="housing",ylab="",main="")

#2.2 credit amount&duration
plot(y=good[,7],x=good[,8],pch=1,
     xlab="duration",ylab="credit amount",main="")
points(y=bad[,7],x=bad[,8],pch=3,
       xlab="duration",ylab="credit amount",main="")

####data processing####
#handle categorical variables
library(nnet)
sex = class.ind(dat[,2])
housing = class.ind(dat[,4])
saving = class.ind(dat[,5])
colnames(saving) = c("s_little","s_moderate","s_quite rich","s_rich")
checking = class.ind(dat[,6])
colnames(checking) = c("c_little","c_moderate","c_rich")
purpose = class.ind(dat[,9])
risk = class.ind(dat[,10])

#correlation plot
datn = dat[,c(-2,-4,-5,-6,-9,-10)]
newdat = cbind(datn,sex,housing,saving,checking,purpose,risk)
mycor = cor(newdat)

library(corrplot)
par(mfrow = c(1,1))
corrplot(mycor,tl.col="black",method="color",shade.lwd = 0.1,tl.cex=0.8,order="hclust")


#4.1 split data by group sampling
newdat = newdat[,-25]
newdat = as.matrix(newdat[order(newdat$good), ])
siz = c(300,700)
n = round(0.7*siz)

library(sampling)
tr = strata(newdat, stratanames='good', size=n, method="srswor")
train = newdat[tr$ID_unit,]
test = newdat[-tr$ID_unit,]
xtr = train[,1:24]
ytr = train[,25]
xte = test[,1:24]
yte = test[,25]

#write.table(train, file ="train data.csv", sep =",")
#write.table(test, file ="test data.csv", sep =",")


#4.2 logistic regression
glm = glm(good~.,data=as.data.frame(train),family=binomial("logit"))

logit.step = step(glm, direction = c("both"))

glm1 = glm(good~s_little+c_little+c_moderate+Duration+female+own,
           data=as.data.frame(train),family=binomial("logit"))

#computing classification rate in training set
pyltr = glm1$fitted.values
yltr = rep(0,length(pyltr))
for(i in 1:length(pyltr)){
  if(pyltr[i]<=0.5){yltr[i]=0}
  else{yltr[i]=1}
}
train_result1 = table(yltr,ytr)
rate1 = (train_result1[1,1]+train_result1[2,2])/sum(train_result1)

#computing classification rate in testing set
y_pred = predict(glm1, as.data.frame(xte))
y_tran = rep(0,length(y_pred))
for(i in 1:length(y_pred)){
  if(y_pred[i]<=0){y_tran[i]=0}
  else{y_tran[i]=1}
}
test_result1 = table(y_tran,yte)
rate2 = (test_result1[1,1]+test_result1[2,2])/sum(test_result1)

library(xgboost)
xgb = xgboost(data = xtr,
              label = ytr,
              eta = 0.1,
              max_depth = 5,
              gamma = 0,
              objective = "binary:logistic",
              nrounds = 15,
              subsample = 0.8,
              colsample_bytree = 0.8,
              seed = 1,
              eval_metric = "error"
)

#computing classification rate in training set
y_pred = predict(xgb, xtr)
y_train = rep(0,length(y_pred))
for(i in 1:length(y_pred)){
  if(y_pred[i]<=0.5){y_train[i]=0}
  else{y_train[i]=1}
}
test_result3 = table(y_train,ytr)
rate3 = (test_result3[1,1]+test_result3[2,2])/sum(test_result3)

#computing classification rate in testing set
y_pred = predict(xgb, xte)
y_test = rep(0,length(y_pred))
for(i in 1:length(y_pred)){
  if(y_pred[i]<=0.5){y_test[i]=0}
  else{y_test[i]=1}
}
test_result4 = table(y_test,yte)
rate4 = (test_result4[1,1]+test_result4[2,2])/sum(test_result4)

paste("classification rate in training set via logit regression is", round(rate1,2))
paste("classification rate in testing set via logit regression is", round(rate2,2))
paste("classification rate in training set via xgboost is", round(rate3,2))
paste("classification rate in testing set via xgboost is", round(rate4,2))