Spatiotemporal Intent Sequence Recommendation on Amap Industrial Dataset 100,000 intent flows (test)
62.7Acc@1GPlan-4B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| GPlan-4BBackbone=Qwen3 4B2026.05 | 62.7 | 0.7759 | |
| GPlan-1.7BBackbone=Qwen3 1.7B2026.05 | 60.5 | 0.745 | |
| TIGERType=generative recommendation model2026.05 | 44.2 | 0.6647 | |
| P5Type=generative recommendation model2026.05 | 43.5 | 0.6581 | |
| ReaRecType=reasoning-based model2026.05 | 42.9 | 0.6442 | |
| BERT4RecType=traditional sequential model2026.05 | 37.4 | 0.5973 | |
| SASRecType=traditional sequential model2026.05 | 36.1 | 0.5865 |