How do I find the best cutoff in roc. how do I set the cutoff for the table function
`roc6<-read.csv("ROC6.csv", header = T)
model6<-glm(Gp~IL6 + CC + Glucose + Lactate + Protein, family = binomial(link = "logit"), data = roc6)
pmodel6<- predict(model6, roc6)
tab6<- table(pmodel6>0.5, roc6$Gp)
tab6
sum(diag(tab6))/sum(tab6)*100
#actual accuracy
table(roc6$Gp)
#model performance
predmodel<- predict(model6)
predictionmodel<- prediction(predmodel, roc6$Gp)
roc6graph<- performance(predictionmodel, "tpr", "fpr")`