Question

enter image description here我对数据科学和R语言还很陌生，所以请耐心等待我:)

我想要的是事后进行逻辑回归模型和混淆矩阵，但是我得到以下警告消息：

glm.fit：发生数字为0或1的拟合概率

我有14个解释变量和1个二进制目标变量用于信用审批（未批准0，已批准1）dataset structure

creditApproval <- read.table(file.choose(), header=TRUE)
summary(creditApproval)
dim(creditApproval)
str(creditApproval)

creditApproval$class <- as.factor(creditApproval$class)

### Setting seed for reproductability
set.seed(2567)

# Normalizing dataset

approvalNORM <- creditApproval 

for(i in 1:length(colnames(creditApproval))-1) {
  if(class(creditApproval[,i]) == "numeric" || class(creditApproval[,i]) == "integer") 
  { minimum<-min(creditApproval[,i])
    maximum<-max(creditApproval[,i]) 
    approvalNORM[,i] <- as.vector(scale(creditApproval[,i],center=minimum,scale=maximum-minimum)) } } 
summary(approvalNORM) # Normalization

# Logistic Regression 

model <- glm(class ~., family=binomial(link='logit'), data=approvalNORM)
summary(model)

glm.fit：拟合的数值出现0或1

0 个答案: