"The most salient pattern is the frequent activation of single-character Hindi tokens, especially those representing common grammatical particles, case markers, and postpositions such as \"क\", \"स\", \"ह\", \"म\", \"ल\", \"न\", \"द\", \"र\", \"ज\", and \"व\". These tokens are highly important for sentence structure and meaning in Hindi, often marking relationships between words, possession, location, agency, and other grammatical functions. Their prominence reflects their central role in the construction and parsing of Hindi sentences."
Score Type | Accuracy | Precision | Recall | F1 score | TPR | TNR | FPR | FNR |
---|---|---|---|---|---|---|---|---|
detection | 0.73 | 1.0 | 0.46 | 0.63 | 0.46 | 1.0 | 0.0 | 0.54 |
fuzz | 0.78 | 0.967 | 0.58 | 0.725 | 0.58 | 0.98 | 0.02 | 0.42 |