Table 2. Comparison of unordered and ordered performance under DSETRec model.
Bold indicates the values where the performance of the ordered and unordered models is relatively better.
| Model | DSETRec (set) | DSETRec (sequence) | Gap | |
|---|---|---|---|---|
| Datasets | Metric (%) | |||
| Yoochoose 1/64 | P@10 | 63.26 | 50.06 | +13.2 |
| MRR@10 | 38.92 | 27.07 | +11.85 | |
| P@20 | 70.70 | 58.89 | +11.81 | |
| MRR@20 | 39.43 | 28.15 | +11.28 | |
| Yoochoose 1/4 | P@10 | 62.88 | 51.66 | +11.22 |
| MRR@10 | 36.84 | 27.38 | +9.46 | |
| P@20 | 69.84 | 60.05 | +9.79 | |
| MRR@20 | 37.33 | 28.23 | +9.1 | |
| Diginetica | P@10 | 39.88 | 38.02 | +1.86 |
| MRR@10 | 21.41 | 18.99 | +2.42 | |
| P@20 | 48.87 | 44.19 | +4.68 | |
| MRR@20 | 21.83 | 16.89 | +4.94 | |