Aggregating Czech-language user comments from Facebook labeled for sentiment classification as part of the Massive Text Embedding Benchmark (MTEB). It focuses on the 't2c' task category across the reviews and written domains to test the efficacy of text embedding models in Slavic languages.
Use Cases
- Evaluate Czech text embedding models using the mteb.get_tasks and mteb.MTEB evaluation pipeline
- Train sentiment classification algorithms on social media text to identify user sentiment in Facebook comments
- Compare the performance of multilingual versus monolingual models on Czech-specific classification tasks
Strengths
- Sourced from Facebook user comments for sentiment classification
- Classified as a 't2c' (text-to-classification) task within the Massive Text Embedding Benchmark
- Focuses on Czech language data within the 'Reviews' and 'Written' domains
- Referenced in ACL anthology W13-1609 for standardized sentiment research