Korean News Articles with Summaries from July 2022

Filled with 24,934 news articles (22,194 train, 2,740 test) with 7 features including document text and human-written summaries. It was created by daekeun-ml via web crawling Naver News for Korean NLP tasks. The articles are from July 1-10, 2022 and focus on IT and economics subjects.

Use Cases

• Train a Korean text summarization model using the document and summary fields.
• Analyze press outlet distribution or writing style differences across the press and category fields.
• Build a news category classifier for IT and economics articles using the title and category columns.
• Study temporal patterns in news coverage during the 10-day period using the date field.

Strengths

• Contains 24,934 total articles, providing a substantial text corpus for model training.
• Includes a dedicated summary field for each article, enabling supervised summarization tasks.
• Covers a specific 10-day period (July 1-10, 2022) for focused temporal analysis.
• Provides a clear train/test split (22,194/2,740 rows) for machine learning evaluation.

Limitations

• Limited temporal scope of only 10 days may not capture seasonal or long-term trends.
• Focus on only two news categories (IT, economics) limits generalizability to other topics.
• Data is from a single news portal (Naver News), which may introduce source-specific bias.

Provenance

Source

Naver News (https://news.naver.com).

Collection Method

Created by the author via web crawling.

Time Range

July 1, 2022 - July 10, 2022.

Freshness

Data covers July 2022; dataset was last updated on the platform in January 2023.

Geography

South Korea (inferred from source and language).