"The highlighted tokens often correspond to morphemes, syllables, or short word fragments that are significant in named entities, place names, or proper nouns, especially in multilingual or transliterated contexts. These fragments frequently appear in the middle or end of words and are commonly found in Indian names, administrative terms, and other culturally specific vocabulary."
Score Type | Accuracy | Precision | Recall | F1 score | TPR | TNR | FPR | FNR |
---|---|---|---|---|---|---|---|---|
detection | 0.73 | 0.829 | 0.58 | 0.682 | 0.58 | 0.88 | 0.12 | 0.42 |
fuzz | 0.55 | 0.532 | 0.82 | 0.646 | 0.82 | 0.28 | 0.72 | 0.18 |