OCLAR is a text corpus of reviews written in Lebanese Arabic. The dataset is intended for sentiment analysis and opinion mining tasks. It was contributed to the UCI Machine Learning Repository.
Use Cases
- Classify review sentiment polarity using labeled text data.
- Train language models to understand dialect-specific vocabulary and sentiment expressions.
- Analyze linguistic features unique to Lebanese Arabic in user-generated content.
- Benchmark sentiment analysis algorithms on a non-Modern Standard Arabic corpus.
Strengths
- Focuses on a specific Arabic dialect (Lebanese), addressing a gap in resources.
- Designed for a core NLP task (sentiment analysis/opinion mining).
- Hosted on the authoritative UCI Machine Learning Repository.
Limitations
- Exact size (row count), column details, and temporal coverage are unknown.
- Potential class imbalance or limited topic diversity within reviews is not specified.
- Data may be temporally stale without a known last update date.
Provenance
- Source
- UCI Machine Learning Repository
- Collection Method
- null
- Time Range
- null
- Freshness
- null
- Geography
- Lebanon