c**********e
发帖数: 2007 | 1 I have two variables (or indicators) of a group of patients,
say A_i and B_i. Each of them (if greater than certain value)
could be an indication of certain disease.
Do I need anything else to construct an ROC curve? |
s*********e
发帖数: 1051 | 2 yes, you need a predicted probability of event = 1. |
D******n
发帖数: 2836 | 3 a variable that indicate the actual situation of the patient,namely whether
he/she has that disease or not.
【在 c**********e 的大作中提到】
: I have two variables (or indicators) of a group of patients,
: say A_i and B_i. Each of them (if greater than certain value)
: could be an indication of certain disease.
: Do I need anything else to construct an ROC curve?
|
c**********e
发帖数: 2007 | 4 Thank you. It is part of logistic procedure. There are macros
roc.sas and rocplot.sas for use. |
