does any one have experience on training binary classifier with skewed dataset?
what's the relationship between tensorflow auc_precision_recall plot and a standard precision-recall plot?
let's say a tensorflow auc_precision_recall like this.
does the highest value is the optimal one?