Low precision of classifiers

[mjuarezm]

Describe the question(问题描述) When I use DeepFM on the ipinyou dataset I obtain a 0.91 AUC, which is good, but when I look at the Precision-Recall curve, the precision-recall trade-offs are very bad. In particular, precision seems to be very low for any decent Recall. In the papers I only see reported the AUC and there's no information about Precision, do you know if this is normal?