"The highlighted tokens are predominantly Korean noun and suffix forms, often ending with \"함\", \"것\", \"음\", \"법\", \"형\", \"율\", \"량\", \"군\", \"객\", \"면\", \"성\", \"전\", \"편\", \"심\", \"득\", \"상\", \"시간\", and similar morphemes. These tokens typically denote abstract concepts, objects, people, or states, and are frequently used as nominalizers or to form compound nouns, reflecting a pattern of marking key semantic units or grammatical roles in Korean text."
Score Type | Accuracy | Precision | Recall | F1 score | TPR | TNR | FPR | FNR |
---|---|---|---|---|---|---|---|---|
detection | 0.55 | 0.857 | 0.12 | 0.211 | 0.12 | 0.98 | 0.02 | 0.88 |
fuzz | 0.61 | 0.923 | 0.24 | 0.381 | 0.24 | 0.98 | 0.02 | 0.76 |