# MDMD
     
MDMD is an open-access, multi-labeled dataset for the Moroccan dialect. It comprises 42,000 comments collected from the social media platform Facebook, with each comment being manually labeled for language classification, sentiment analysis, and offensive language detection.

Each comment in our dataset has been meticulously annotated with three distinct labels, providing valuable contextual information for subsequent analyses.

### The first label focuses on the language used in the comment, allowing us to explore the multilingual nature of user interactions through seven distinct labels:
      DAL (Dialect Latin Alphabet)
      DAA (Dialect Arabic Alphabet)
      ARC (Classic Arabic)
      FRN (French)
      ANG (English)
      ESP (Spanish)
      AUT (Other/Undetermined)
### The second label facilitates sentiment analysis, a crucial aspect of understanding user attitudes and emotions. This label helps determine whether a comment conveys positive, negative, or neutral sentiments:
      1: Positive
      2: Negative
      3: Neutral
### The third label plays a critical role in detecting potentially offensive language within comments. We classify the comments as offensive (Offen.) or non-offensive (non-Offen.):
      0: non-Offen.
      1: Offen.