Hello group, I am trying to compare the classification accuracies of multiple classifiers using ROC curves. I have created multiple ROC plots for each of these classifiers but I am looking for a single graph where I can overlay each of these individual ROC plots. Does anyone know anything about creating such overlapping ROC plots in one graph? Below is the sample code I have used: Classifier 1: require (e1071) require(ROCR) train=read.table('C:/train.txt', header=T, sep='\t') y=as.factor(train[,2]) x=train[,3:5] require(e1071) model=svm(x,y,type='C', kernel='linear', cross=10) test=read.table('C:/test.txt', header=T, sep='\t') test_x=test[,3:5] pred=predict(model,data.frame(test_x)) pred1=prediction(as.numeric(pred),test[,2]) perf=performance(pred1,'tpr','fpr') plot(perf, col='red') Classifier 2: ... ... Classifier 3: ... ... Now, I want to combine all these plots in one so that I can better compare. Thanks, Yadav [[alternative HTML version deleted]]