UD_Manx-Cadhan
Treebank card from UD release 2.18: metadata and texts come from UD_Manx-Cadhan/README.md; size is counted by streaming the .conllu files and checked against the release stats.xml. Generated by tools/treebanks build; do not edit by hand. Papers: Papers: UD_Manx-Cadhan; seeds: seeds/; converter: convert.md.
| Field | Value |
|---|---|
| UD name | UD_Manx-Cadhan |
| Language | Manx (gv) |
| Family | IE |
| Treebank code | gv_cadhan |
| UD page | universaldependencies.org/treebanks/gv cadhan/index.html |
| License | CC BY-SA 4.0 |
| Genres | news, fiction, nonfiction, blog, social, wiki, web, bible |
| Text in data | yes |
| Parallel | no |
| In UD since | UD v2.7 |
| Authors | Kevin Scannell |
| Contact | kscanne@gmail.com |
| Development | outside the UD repo (elsewhere): fixes go to the source or the converter |
Annotation layers
| Layer | Origin |
|---|---|
| Lemmas | manual native |
| UPOS | manual native |
| XPOS | manual native |
| Features | manual native |
| Relations | manual native |
UD scale: manual native — annotated by hand directly in UD; converted from manual — converted from manual annotation in another scheme; converted with corrections — converted and partly corrected by hand; automatic with corrections — automatic, with partial manual corrections; automatic — automatic; not available — the layer is absent.
Size
| Split | Sentences | Tokens | Words | Multiword tokens | Empty nodes |
|---|---|---|---|---|---|
| train | 1 172 | 9 325 | 10 378 | 1 030 | 0 |
| test | 1 164 | 9 255 | 10 347 | 1 066 | 0 |
| total | 2 336 | 18 580 | 20 725 | 2 096 | 0 |
Tokens are surface tokens (Spanish vámonos is one), words are syntactic words (vamos + nos); multiword tokens are those split into several words; empty nodes are the N.M nodes of the enhanced graph. Splits dev are absent from the release.
Summary
This is the Cadhan Aonair UD treebank for Manx Gaelic, created by Kevin Scannell.
Introduction
The sentences in the original treebank (UD v2.7) came from a large web-crawled corpus of Manx consisting of more than 8M words of text. This corpus was segmented by sentences, shuffled, and then 300 random sentences were chosen to be tagged. I believe that the web corpus is nearly comprehensive, which is to say it contains virtually all non-trivial Manx language texts on the open web. Therefore, the initial release of the UD corpus was as close as possible to a “random sample” of Manx on the web. As such, the Bible is heavily represented, but there are also quite a few sentences from modern sources such as the Manx Wikipedia, news stories from Manx Radio, blog posts, translations of literature, etc.
All POS tags and dependency annotations in the initial release were added manually. The upstream source code and datasets supporting the treebank can be found here.
The treebank was expanded greatly for UD v2.8, released in May 2021. Most of the new sentences in this release came from Phil Kelly's Manx "usage file", taken from example sentences in various Manx-English dictionaries. Full morphological features were added to this release as well.
Only a test file was provided for releases 2.7 through 2.10. With v2.11 (November 2022), we randomly resplit the corpus into test and train files with about 10k tokens each, following the recommendation of the UD maintainers. This means that any experiments run on earlier versions of the treebank would need to be rerun with the new split.
This page in Markdown → https://md.mova.elamur.ai/free/en/docs/treebanks/gv/cadhan/