Fig. 9
From: Joint embedding–classifier learning for interpretable collaborative filtering
Validation metrics for JELI with different order m of the underlying RHOFM on a “second-order” synthetic data set across 100 iterations (random seeds for splitting between training and testing data sets)