# How to get the area under the curve from multivariate analysis

I have two groups, A and B. They have a number of variables as follows

  Group, Age, Sex, NumberofSomething, NumberofSomething2


I want to determine which of the variables and combination of variables best predicts membership of a group. I then want to calculate an area under the ROC curve for this.

My question is: Does the multivariate analysis also tell me the best combination of factors (is this done by linear discriminant analysis in my case?), and also, how do I then use the results of the multivariate analysis to generate the area under the ROC curve?

• "Multivariate" means multiple response variables while "multivariable" indicates multiple predictors. Do you have multiple groups which people are members of? – Hugh Nov 13 '16 at 14:29
• I have two groups only – Sebastian Zeki Nov 13 '16 at 14:36
• Can people be a member of neither group? – Hugh Nov 13 '16 at 14:37
• No. You are either one or the other – Sebastian Zeki Nov 13 '16 at 14:40