3,000 balanced sentences written in Tunisian Arabizi, a script combining Latin characters and numerals. The data is preprocessed and labeled with binary sentiment classes across various social media topics.
Use Cases
- Train sentiment analysis models specifically for Tunisian dialectal Arabic using the positive and negative labels
- Develop transliteration or normalization tools to convert Tunisian Arabizi (Latin script + numerals) into standard Arabic script
- Perform linguistic analysis on the usage of numerals as phonetic substitutes in North African social media communication
Strengths
- 3,000 balanced sentences annotated for sentiment analysis
- Focuses on Tunisian Arabizi, which utilizes numerals as substitutes for Arabic characters within a Latin script framework
- Includes binary sentiment labels categorized as positive and negative
- Covers diverse topics sourced from social media platforms