Maybe a silly question, but need your advice and comments. I have data entries (compounds) with two values, Value1 and Value2 (activity for two proteins). Is it meaningful to build the ROC curve for these data entries with class1 (zero) as value1 and class2 (one) as value2?
So, if the ROC AUC is close to 0.5 - means that we cannot distinguish between the Class1 and Class2.
Are those 2 proposals correct? Thank's a lot!