Sentiment Analysis on NSMC (test)
0.894AccuracyBERT
Evaluation results
| Method | Links | |
|---|---|---|
| BERTTokenization=MorSubword, Vocab Size=32K2026.04 | 0.894 | |
| BERTTokenization=Subword, Vocab Size=32K2026.04 | 0.8891 | |
| BERTTokenization=Character, Vocab Size=2K2026.04 | 0.8889 | |
| KOMBO_JamoTokenization=Jamo, Vocab Size=1702026.04 | 0.887 | |
| KOMBO_CjiTokenization=Cji, Vocab Size=1362026.04 | 0.8808 | |
| BERTTokenization=Jamo, Vocab Size=1702026.04 | 0.8778 | |
| KOMBO_StrokeTokenization=Stroke, Vocab Size=1302026.04 | 0.8747 | |
| BERTTokenization=Morpheme, Vocab Size=32K2026.04 | 0.874 | |
| BERTTokenization=Stroke, Vocab Size=1302026.04 | 0.8684 | |
| BERTTokenization=Cji, Vocab Size=1362026.04 | 0.867 | |
| KOMBO_BTSTokenization=BTS, Vocab Size=1122026.04 | 0.864 | |
| BERTTokenization=BTS, Vocab Size=1122026.04 | 0.8623 | |
| BERTTokenization=Word, Vocab Size=64K2026.04 | 0.7416 |