Document Classification on EC News ZH
77.85AccuracyNMTM + LLM-XTM
Evaluation results
| Method | Links | |
|---|---|---|
| NMTM + LLM-XTMEvaluation Protocol=Intra-lingual (-I), Features=document-topic vectors (θ), Classifier=Linear SVM2026.05 | 77.85 | |
| NMTMEvaluation Protocol=Intra-lingual (-I), Features=document-topic vectors (θ), Classifier=Linear SVM2026.05 | 77.42 | |
| InfoCTM + LLM-XTMEvaluation Protocol=Intra-lingual (-I), Features=document-topic vectors (θ), Classifier=Linear SVM2026.05 | 76.82 | |
| InfoCTMEvaluation Protocol=Intra-lingual (-I), Features=document-topic vectors (θ), Classifier=Linear SVM2026.05 | 76.52 | |
| InfoCTM + LLM-XTMEvaluation Protocol=Cross-lingual (-C), Features=document-topic vectors (θ), Classifier=Linear SVM2026.05 | 58.75 | |
| InfoCTMEvaluation Protocol=Cross-lingual (-C), Features=document-topic vectors (θ), Classifier=Linear SVM2026.05 | 55.92 | |
| NMTM + LLM-XTMEvaluation Protocol=Cross-lingual (-C), Features=document-topic vectors (θ), Classifier=Linear SVM2026.05 | 53.33 | |
| NMTMEvaluation Protocol=Cross-lingual (-C), Features=document-topic vectors (θ), Classifier=Linear SVM2026.05 | 51.39 |