ACTib Part I: Annotated Classical Tibetan Corpus

The Annotated Corpus of Classical Tibetan (ACTib), Part I is a part-of-speech tagged version of the Buddhist Digital Resource Center's digitized Tibetan etext collection. It was created using a memory-based tagger trained on a separate POS-tagged corpus of Classical Tibetan. The corpus includes files that were not manually corrected and contains some annotations from corrupted source files.

Use Cases

• [Part-of-speech tagging model training] based on [the annotated training data from Hill & Garrett]
• [Linguistic analysis of Classical Tibetan] based on [the POS-tagged corpus]
• [Evaluation of automated text processing tools] based on [the uncorrected, machine-generated annotations]
• [Text mining of Buddhist literature] based on [the BDRC digitized etext collection]

Strengths

• Derived from a large, established collection of digitized Tibetan texts compiled by the Buddhist Digital Resource Center
• Utilizes a specific, documented training corpus and tagging methodology (Memory-Based Tagger/TiMBL)

Limitations

• Files are not post-processed or manually corrected, indicating potential annotation errors
• A small number of files were annotated from corrupted source XML, which may affect data integrity
• Key metadata such as row count, size, and specific column structure is unavailable

Provenance

Source

Buddhist Digital Resource Center (BDRC) digitized text collection

Collection Method

Automated part-of-speech tagging using a memory-based tagger (TiMBL) trained on a separate annotated corpus