NamSyntax/vietnamese-news-classification

Vietnamese News Classification Dataset (1.3M)

Dataset Description

This dataset contains approximately 1.3 million Vietnamese news articles collected from major online news portals. Structured similarly to the popular AG News dataset, it serves as a valuable resource for experimenting with multi-class text classification in Vietnamese.

The dataset covers 11 topics (categories) ranging from current affairs, sports, technology, to entertainment.

• —Curated by: Nam Syntax
• —Language: Vietnamese
• —Total Rows: ~1.3M
• —Task: Multi-class Text Classification

Dataset Structure

Data Fields

Each instance in the dataset consists of the following fields:

• —label: The category label id (Integer, 0-10).
• —title: The title of the news article (String).
• —description: A short summary/abstract of the article (String).
• —text: The combination of title and description (and body content if available) used for training (String).

Example

{
"label": 0
"title": "Ba xe tông liên hoàn trên đường tránh Quảng Bình",
"description": "Xe khách bị rách hông bên phải sau cú tông vào xe tải chạy ngược chiều...",
"text": "Ba xe tông liên hoàn trên đường tránh Quảng Bình. Xe khách bị rách hông...",
}