ThaiTrees

A 342M-token corpus of Thai drawn from news, Wikipedia, spoken transcripts and social media, automatically parsed under the Universal Dependencies framework. It is released as three artefacts: a raw text corpus, a frequency lexicon, and a dependency-parsed corpus in CoNLL-U.

Dataset Summary

ThaiTrees contains 341,967,133 tokens across 366,120 documents in four domains (news, Wikipedia, spoken transcripts, social media). Document identifiers are shared across all three artefacts, so results can be joined at the document level. Token identifiers do not align across the lexicon and the parsed corpus, which index different tokenisations of the same raw text (see Processing Pipeline).

• Raw text corpus — 366,120 documents and 341,967,133 AttaCut-segmented tokens, released as Parquet.
• Frequency lexicon — 451,252 unique word forms and 255,434,917 filtered tokens, with per-domain count, rank, frequency per million, document frequency and IDF. Covers the full raw corpus.
• Dependency-parsed corpus — 199,836,464 dependency edges over 203,892,200 tokens in 10-column CoNLL-U, with PhayaThaiBERT POS tags.

Thai already has a manually annotated dependency treebank, UD Thai-TUD (3,627 dependency trees, 77,215 tokens). ThaiTrees complements rather than replaces it: ThaiTrees is automatically parsed and not validated against human annotation, so use UD Thai-TUD as the gold standard for parser evaluation.

Languages

Thai (th).

Dataset Structure

Configurations

Config	Description	Splits
raw	Raw text corpus	news, wikipedia, spoken, social_media
lexicon	Frequency lexicon	train (single split)
parsed	Dependency-parsed CoNLL-U	news, wikipedia, spoken, social_media

Data Fields

raw configuration (one row per document):

Field	Type	Description
domain	string	One of news / wikipedia / spoken / social_media
id	string	Document identifier (e.g. news_1); matches doc_id in parsed
url	string	Source URL
title	string	Document title
categories	list[string]	Source categories or tags, where available (may be null)
text	string	The original, unsegmented document text
segmented_text	string	The document text, AttaCut-tokenised, with | between tokens
token_count	int64	Number of AttaCut tokens in the document

lexicon configuration (one row per unique word form, 34 columns):

Field	Type	Description
word	string	The word form
total_count	int	Cross-corpus occurrence count
total_rank	int	Cross-corpus rank by frequency
total_freq_per_million	float	Frequency per million tokens, cross-corpus
total_doc_count	int	Number of documents (corpus-wide) containing this word
{domain}_count	int	Occurrence count in {domain}
{domain}_rank	int	Within-domain rank
{domain}_freq_per_million	float	Frequency per million tokens within domain
{domain}_doc_count	int	Number of documents in {domain} containing this word
{domain}_doc_freq	float	Document frequency in {domain} (0–1)
{domain}_idf	float	IDF computed over {domain} documents
domain_count	int	How many of the four domains the word appears in
is_cross_domain	bool	True if the word appears in all four domains
word_length	int	Character length of the word
idf	float	Cross-corpus IDF
max_domain	string	The domain in which the word's freq/M is highest

A SQLite copy of the lexicon (lexicon/lexicon.sqlite) is also provided.

parsed configuration (one row per sentence):

Field	Type	Description
doc_id	string	Document identifier (matches id in raw)
domain	string	Domain tag
sent_id	uint32	Sentence index within the document
tokens	list[string]	FORM (CoNLL-U column 2)
lemma	list[string]	LEMMA (column 3); placeholder, always .
upos	list[string]	UPOS (column 4); PhayaThaiBERT POS tags
xpos	list[string]	XPOS (column 5); placeholder, always .
feats	list[string]	FEATS (column 6); placeholder, always _
head	list[int32]	HEAD (column 7); 1-indexed, 0 = root
deprel	list[string]	DEPREL (column 8); Universal Dependencies relation
deps	list[string]	DEPS (column 9); placeholder, always _
misc	list[string]	MISC (column 10); character offsets

Data Splits

Split	Documents	Tokens	Sentences
news	100,846	52,641,127	897,348
wikipedia	175,069	101,378,583	1,303,443
spoken	2,505	30,238,809	266,944
social_media	87,700	157,708,614	2,052,040
Total	366,120	341,967,133	4,519,775

The parsed corpus contains 203,892,200 tokens and 199,836,464 dependency edges. The lexicon config has a single train split with 451,252 rows (one per unique word form; see Processing Pipeline).

Dataset Creation

Curation Rationale

Studying syntactic patterns in naturally occurring language requires a large parsed corpus, but manual annotation is costly and difficult to scale. Thai has a manually annotated dependency treebank (UD Thai-TUD) for training and evaluating parsers, but lacked a large automatically parsed corpus for quantitative syntactic research. ThaiTrees fills that gap: a four-domain corpus whose sentences are parsed automatically under Universal Dependencies, released with a frequency lexicon in machine-readable formats.

Source Data

Domain	Source	Collection
news	ThaiPBS public-broadcaster website	Articles publicly available on the website in December 2025
wikipedia	Thai Wikipedia	MediaWiki random-article endpoint, namespace 0, batches of 500, disambiguation pages excluded, reference sections truncated; sampled December 2025
spoken	YouTube: Thai PBS Podcast network, independent podcasts (bigboung, BeSider), film subtitles, Prachatai political-commentary livestreams, SaltymanTH livestreams	youtube-transcript.io API; channels selected manually for their manual transcription
social_media	Wisesight sentiment dataset and Pantip forum	The same sources used in training WangchanBERTa and PhayaThaiBERT

Processing Pipeline

The lexicon and the parsed corpus are produced in separate branches that apply word segmentation to different units of text: whole documents for the lexicon, individual sentences for parsing.

Frequency lexicon. AttaCut (Chormai et al., COLING 2020), a CNN-based Thai word segmenter, segments each whole document. Lexicon construction then removes punctuation, pure-numeric tokens, emoji, whitespace-only tokens, and forms whose total corpus frequency is at most five. After filtering, 255,434,917 tokens remain across 451,252 unique forms; 39,465 forms (8.7%) appear in all four domains.

Dependency-parsed corpus.

1. Sentence preparation. The raw text has newlines and runs of spaces collapsed to a single space, and social-media documents are capped at 500,000 characters (this truncates 149 long, concatenated documents).
2. Sentence segmentation with PyThaiNLP CRFcut, followed by AttaCut re-applied to each sentence string.
3. Dependency parsing with AttaParse 1.0, a Thai-specific wrapper around Stanza, on the pre-tokenised input. Whitespace-only tokens are discarded at the parser's pre-tokenised entry point. One deterministic parser failure on Wikipedia chunk 16 is excluded from the release.
4. POS tagging with a PhayaThaiBERT model fine-tuned on UD Thai-TUD (nlp-chula/phayathaibert-thai-pos-tagger), which achieves 90.64% accuracy and 81.34% macro F1 on the UD Thai-TUD test set. Its predictions overwrite the UPOS column (column 4) of the CoNLL-U output.

Additional Information

Dataset Curators

Attapol T. Rutherford and Papatchol Thientong, Department of Linguistics, Chulalongkorn University. Code: https://github.com/nlp-chula/thaitrees.

Citation Information

@misc{rutherford2026thaitreesthaisyntacticdependency,
      title={ThaiTrees: Thai Syntactic Dependency Trees Across Domains}, 
      author={Attapol T. Rutherford and Papatchol Thientong},
      year={2026},
      eprint={2609.27558},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2609.27558}, 
}

Loading Examples

from datasets import load_dataset

# Raw text, per-domain split
ds_news = load_dataset("nlp-chula/ThaiTrees", "raw", split="news")
print(ds_news[0])
# {'domain': 'news', 'id': 'news_1', 'url': 'https://www.thaipbs.or.th/news/content/1',
#  'title': 'นักวิจัยออสเตรเลียเผยสาเหตุฉลามโจมตีมนุษย์', ...,
#  'segmented_text': 'นัก|วิจัย|ออสเตรเลีย|เผย|...', 'token_count': 230}

# Frequency lexicon (single split)
ds_lex = load_dataset("nlp-chula/ThaiTrees", "lexicon", split="train")
print(ds_lex[0])
# {'word': 'ที่', 'total_count': 6300593, 'total_rank': 1, ...}

# Dependency-parsed CoNLL-U, per-domain split
ds_parsed = load_dataset("nlp-chula/ThaiTrees", "parsed", split="news")
print(ds_parsed[0])
# {'doc_id': 'news_5028', 'domain': 'news', 'sent_id': 0,
#  'tokens': ['ย้าย', 'ผกก.', 'มีนบุรี', '-', 'หนองจอก', ...],
#  'upos': ['VERB', 'NOUN', 'PROPN', 'PUNCT', 'PROPN', ...], ...}