"The highlighted tokens are primarily Hindi morphemes, suffixes, and inflections, as well as common connective words and noun/adjective endings. There is frequent emphasis on grammatical markers, case endings, and conjunct forms, along with some corrupted or non-standard characters. The pattern reflects a focus on the morphological structure and syntactic connectors in Hindi text, especially those that signal relationships between words or modify meaning."
Score Type | Accuracy | Precision | Recall | F1 score | TPR | TNR | FPR | FNR |
---|---|---|---|---|---|---|---|---|
detection | 0.8 | 1.0 | 0.6 | 0.75 | 0.6 | 1.0 | 0.0 | 0.4 |
fuzz | 0.83 | 0.946 | 0.7 | 0.805 | 0.7 | 0.96 | 0.04 | 0.3 |