Cross-lingual Topic Modeling on Rakuten Amazon
0.04CNPMIInfoCTM + LLM-XTM
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| InfoCTM + LLM-XTMIntegrated Enhancement=LLM-XTM2026.05 | 0.04 | 0.87 | 0.035 | |
| XTRA + LLM-XTMIntegrated Enhancement=LLM-XTM2026.05 | 0.037 | 0.945 | 0.035 | |
| XTRA2026.05 | 0.034 | 0.966 | 0.033 | |
| InfoCTM2026.05 | 0.032 | 0.87 | 0.028 | |
| u-SVD2026.05 | 0.027 | 0.571 | 0.015 | |
| SVD-LR2026.05 | 0.026 | 0.558 | 0.015 | |
| MCTA2026.05 | 0.021 | 0.272 | 0.006 | |
| NMTM + LLM-XTMIntegrated Enhancement=LLM-XTM2026.05 | 0.016 | 0.666 | 0.011 | |
| NMTM2026.05 | 0.012 | 0.633 | 0.008 | |
| MTAnchor2026.05 | 0.001 | 0.214 | 0 |