A large collection of manually annotated English news articles categorized for target-dependent sentiment classification. It provides human-verified sentiment labels associated with specific targets mentioned within news text.
Use Cases
- Train target-dependent sentiment classification models using the manual sentiment labels
- Evaluate entity-specific sentiment analysis performance on news-domain text
- Analyze sentiment polarity towards specific targets within news sentences
Strengths
- Manually annotated sentiment labels for target-dependent classification
- Focuses on English-language news article content
- Includes target-specific sentiment indicators for entity-level analysis