200,000 French-language movie reviews from Allociné.fr categorized into 100,000 positive and 100,000 negative sentiment labels. The collection spans user reviews from 2006 to 2020 and is structured into training, validation, and test splits.
Use Cases
- Train a text classification model to distinguish between positive and negative sentiment using the review text and labels
- Fine-tune French-language transformers on movie-specific terminology and slang found in the review text
- Benchmark sentiment analysis algorithms using the predefined 20,000-sample test set
Strengths
- 200,000 labeled reviews with a perfect 50/50 balance between positive and negative classes
- Data collection period covering 14 years of community reviews from 2006 to 2020
- Standardized splits consisting of 160,000 training, 20,000 validation, and 20,000 test examples