2,909,551 news articles from the SogouCA and SogouCS corpora categorized into 5 classes. The dataset features 450,000 training and 60,000 testing samples where original Chinese characters are converted to Pinyin.
Use Cases
- Train a text classification model to predict one of 5 categories using Pinyin-encoded news content
- Benchmark phonetic-to-semantic mapping by analyzing the relationship between Pinyin strings and news domains
- Develop neural networks for sequence classification on Pinyin-based text data
Strengths
- 2,909,551 news articles sourced from SogouCA and SogouCS corpora
- 5 distinct news categories labeled based on URL domain names
- Text content is entirely converted from Chinese characters to Pinyin
- Structured with 90,000 training and 12,000 testing samples per category