"The highlighted tokens are primarily Bulgarian morphemes or word segments, especially those forming relative pronouns (such as \"която\", \"който\", \"което\", \"кои\"), as well as common suffixes and inflections (like \"ъ\", \"оя\", \"ой\", \"г\", \"точ\", \"еч\", \"еств\", \"ал\", \"ото\"). These segments often appear at the boundaries of words or as part of grammatical constructions, indicating a focus on morphological and syntactic markers in Bulgarian text."
Score Type | Accuracy | Precision | Recall | F1 score | TPR | TNR | FPR | FNR |
---|---|---|---|---|---|---|---|---|
detection | 0.73 | 0.96 | 0.48 | 0.64 | 0.48 | 0.98 | 0.02 | 0.52 |
fuzz | 0.71 | 0.957 | 0.44 | 0.603 | 0.44 | 0.98 | 0.02 | 0.56 |