Web Data Extraction on SWDE (F1 Score)
94.78F1 ScoreCo-Scraper
Evaluation Results
| Method | Links | |
|---|---|---|
| Co-ScraperEvaluation domain=Univ. (fully out-of-domain)2026.06 | 94.78 | |
| WebFormerNumber of seed sites for training=12026.06 | 86.58 | |
| MarkupLMBASENumber of seed sites for training=12026.06 | 84.31 | |
| Render-FullNumber of seed sites for training=12026.06 | 84.3 | |
| SimpDOMNumber of seed sites for training=12026.06 | 83.06 | |
| FreeDOMNumber of seed sites for training=12026.06 | 82.32 | |
| Pruned HTMLgt + Qwen3-8BEvaluation domain=Univ.2026.06 | 67.67 | |
| Raw HTML + Qwen3-8BEvaluation domain=Univ.2026.06 | 58.04 | |
| Pruned HTMLgt + Qwen3-8BEvaluation domain=Non-Univ.2026.06 | 54.59 | |
| Pruned HTMLbase + Qwen3-8BEvaluation domain=Non-Univ.2026.06 | 53.57 | |
| Pruned HTMLbase + Qwen3-8BEvaluation domain=Univ.2026.06 | 52.58 | |
| Raw HTML + Qwen3-8BEvaluation domain=Non-Univ.2026.06 | 48.31 |