画ROC曲线

R软件：如何画ROC曲线

> install.packages("pROC")   安装pROC包

> install.packages("ggplot2")安装ggplot2包

library(pROC) #调用pROC包

library(ggplot2)  #安装ggplot包

 data(aSAH)# 引入案例数据

View(aSAH) #展示案例数据

roca<-roc(aSAH$outcome,aSAH$s100b)
rocb<-roc(aSAH$outcome,aSAH$wfns)
rocc<-roc(aSAH$outcome,aSAH$ndka)
auc(roca)
auc(rocb)
auc(rocc)

> auc(roca)
Area under the curve: 0.7314
> auc(rocb)
Area under the curve: 0.8237
> auc(rocc)
Area under the curve: 0.612
> 

 plot(roca)

plot(rocb)

#ROC曲线下面积的比较
roc.test(roca,rocb,method = "delong") 

        DeLong's test for two correlated ROC curves

data:  roca and rocb
Z = -2.209, p-value = 0.02718
alternative hypothesis: true difference in AUC is not equal to 0
sample estimates:
AUC of roc1 AUC of roc2 
  0.7313686   0.8236789 

g3<-ggroc(list(s100b=roca,wfns=rocb,ndka=rocc))
g3