CTR prediction on Dianping-Food (test)
0.85AUCKGNN-LS
Evaluation results
| Method | Links | |
|---|---|---|
| KGNN-LS2019.05 | 0.85 | |
| LibFM + TransETransE dimension=322019.05 | 0.839 | |
| SVD2019.05 | 0.838 | |
| LibFMdimension={1, 1, 8}, training epochs=502019.05 | 0.837 | |
| RippleNetd=32, H=2, lambda1=10^-7, lambda2=0.02, eta=0.012019.05 | 0.833 | |
| CKEembedding dimension=64, KG training weight=0.12019.05 | 0.802 | |
| PERmeta-paths=user-restaurant-dish-restaurant, user-restaurant-business_area-restaurant, user-restaurant-tag-restaurant2019.05 | 0.746 |