Uzbek News Articles for Zero-Shot Text Classification

10 categories, including Politics, Economy, and Technology, classify Uzbek news texts. The dataset was created by risqaliyevds and published on Hugging Face in June 2024. It is structured in JSON format for zero-shot classification tasks.

Use Cases

• Training zero-shot classifiers to assign categories like 'Politics' or 'Sports' to Uzbek news articles.
• Benchmarking cross-lingual transfer learning models using the predefined 'classes' field for label mapping.
• Fine-tuning multilingual language models for text categorization in the 'Foreign News' or 'Ecology' domains.
• Analyzing topic distribution across Uzbek media sources based on the provided news categories.

Strengths

• Covers 10 distinct news categories for classification.
• Data sourced from real-world Uzbek news websites, providing authentic text.

Limitations

• Unknown total number of text samples (row count).
• Potential class imbalance or geographic bias due to unspecified news source selection.
• Lacks detailed column descriptions beyond the 'classes' structure.

Provenance

Source

Hugging Face dataset uploaded by author 'risqaliyevds'.

Collection Method

Prepared based on content from Uzbek news websites.

Freshness

Last updated on June 6, 2024.

Geography

Uzbekistan (implied by language and news focus).