I am still a rookie. I am training a random forest model and I am getting 0.27 for area under PR and 0.85 for area under ROC. My data was very imbalanced for the negative labels and I did perform a down-sampling to balance it. I understand what ROC and PR are, but what I am not getting is how to interpret a low area under area PR vs. high area under ROC.
Asked
Active
Viewed 8 times
