100 text examples from the AG News corpus, curated for classification tasks. The dataset was created by TimeRobber and published on Hugging Face in June 2022.
Use Cases
- Train a text classifier to predict news categories using the article text field.
- Benchmark model performance on a small, curated subset of the AG News corpus.
- Fine-tune a pre-trained language model for topic classification on news headlines and snippets.
- Analyze feature extraction methods for short news text samples.
Strengths
- Derived from the established AG News corpus, providing a known benchmark.
- Contains 100 curated examples for focused experimentation.
Limitations
- Very small sample size of 100 rows limits statistical power and generalizability.
- No column details provided, obscuring the exact structure and available labels.
- Data is from 2022 and may not reflect current news topics or language.
Provenance
- Source
- Hugging Face, user TimeRobber.
- Collection Method
- Subset of the AG News corpus.
- Time Range
- null
- Freshness
- Last updated in June 2022.
- Geography
- null