TUNIZI

该数据集名为TUNIZI，是一个情感分析数据集，包含了超过9,000条以突尼斯阿拉伯字母书写的突尼斯句子。这些句子是从YouTube社交网络评论中收集而来，覆盖了各种主题。该数据集由五位突尼斯母语者标注，其中47%为积极评论，53%为负面评论，保持了良好的平衡性。规模上，该数据集包含了超过9,000个句子，其任务专注于情感分析。

Named TUNIZI, this is a sentiment analysis dataset containing over 9,000 Tunisian sentences written in Tunisian Arabic script. These sentences were collected from YouTube social media comments and cover a wide variety of topics. Annotated by five native Tunisian speakers, the dataset features a well-balanced class distribution, with 47% positive comments and 53% negative comments. With a scale of over 9,000 sentences, TUNIZI is dedicated to sentiment analysis tasks.