| Step | Description | |------|-------------| | | Find potential boundaries using punctuation ( . , । , ! , ? , ॥ ) or line breaks | | 2. Abbreviation handling | Protect known abbreviations (e.g., "Dr." , "v. l." , "śrī." ) from splitting | | 3. Context analysis | Check capitalization (if available) and surrounding word classes | | 4. Machine learning (optional) | In advanced versions, a binary classifier decides boundary vs. non-boundary |
pip install plic-sbd
To adapt plic-sbd to a new language/domain: plic-sbd