Korean News and Commentary Discourse Comprehension Dataset

#discourse comprehension #event narrative understanding #cross-sentence relation analysis #Korean news comprehension #Korean discourse analysis #formal Korean language learning

5000
0.5GB
JSON
CATL
MOBIUSI INC

Updated:2026-11-05

AI Analysis & Value Prop

This dataset contains news articles, columns, and editorial-style documents in Korean suited to intermediate and advanced learners. JSON annotations capture event structures, arguments and evidence, discourse and causal relations, stance, attribution sources, and document summaries. The collection focuses on cross-sentence meaning and argument organization in formal written Korean, supporting domain-specific SFT as well as training and evaluation for document understanding, information extraction, and discourse analysis.

Dataset Insights

Sample Examples

{}

Technical Specifications

Field	Type	Description
file_name	string	File name
event_structure	array	Structured information extracted from the document about events, including triggers, participants, time, and location.
source_document	object	The original Korean news article, column, or editorial-style document to be analyzed, together with basic metadata.
stance_analysis	object	Analysis of the attitudes and stance strength expressed by the author, media outlet, or other entities toward issues, events, and claims.
causal_relations	array	Causes, outcomes, causal trigger expressions, and relation types extracted from the document.
document_summary	object	A comprehensive summary of the document's topic, main content, and key viewpoints.
argument_structure	array	Annotations of the document's central claims, supporting evidence, reasoning relations, and any counterarguments.
attribution_sources	array	Annotations of people, organizations, and other entities quoted, paraphrased, or cited as information sources in the document.
discourse_relations	array	Annotations of connections between sentences or paragraphs that capture how formal written text is organized.

Compliance Statement

Authorization Type	Proprietary - Commercial AI Training License (No Redistribution)
Commercial Use	Requires exclusive subscription or authorization contract (monthly or per-invocation charging)
Privacy and Anonymization	No PII, no real company names, simulated scenarios follow industry standards
Compliance System	Compliant with China's Data Security Law / EU GDPR / supports enterprise data access logs

Frequently Asked Questions

What types of Korean texts are included?: The collection focuses on news articles, columns, and editorial-style documents suited to intermediate and advanced Korean learners, with an emphasis on formal written language, cross-sentence meaning, and argument structure.

What annotations and data format are provided?: The data uses a JSON structure with fields for source documents, event structures, argument structures, discourse relations, causal relations, stance analysis, attribution sources, and document summaries. Refer to the product page for sample records and detailed field specifications.

Which training and evaluation tasks does it support?: It can support Korean document understanding, event extraction, argument structure identification, causal relation extraction, stance analysis, and discourse relation identification. It is also useful for evaluating cross-sentence comprehension of formal Korean text.

How can I obtain the data, and what are the license terms?: Access options, available versions, and licensing depend on the terms shown on the product page. Before use, confirm that the license covers your intended application and comply with any copyright requirements that apply to the source texts and annotations.

Can't find the data you need?

Post a request and let data providers reach out to you.

Get this Dataset

Verified for Enterprise Use

Cite this Work

@dataset{Mobiusi2026,
  title={Korean News and Commentary Discourse Comprehension Dataset},
  author={MOBIUSI INC},
  year={2026},
  url={https://www.mobiusi.com/datasets/baa7389b808f8d9d3f66b46e26004dc1},
  urldate={2026-10-16},
  keywords={Korean news dataset,Korean discourse comprehension,Korean reading comprehension data,Korean event extraction,argument structure annotation,Korean causal relation detection,Korean stance analysis,Korean document understanding,Korean SFT training},
  version={1.0}
}

Using this in research? Please cite us.

placeholder

Scene

Domain

Modality

Task