misyaz/BAREC-Corpus-v1.0

BAREC（平衡阿拉伯语可读性评估语料库）是一个用于细粒度阿拉伯语可读性评估的大规模数据集。该数据集包含超过100万个单词，在句子级别上标注了19个可读性级别，并额外映射到7、5和3个级别的更粗粒度方案中。数据集支持多类可读性分类任务，包括19级、7级、5级和3级分类。数据实例包含句子文本、单词计数、词法信息、可读性级别（19级、7级、5级和3级）、标注者ID、文档来源、书籍、作者、领域（艺术与人文、STEM或社会科学）和文本类别（基础、高级或专业）。数据集分为训练集（80%）、开发集（10%）和测试集（10%），并在可读性级别、领域和文本类别上保持平衡。

BAREC (the Balanced Arabic Readability Evaluation Corpus) is a large-scale dataset for fine-grained Arabic readability assessment. The dataset includes over 1M words, annotated at the sentence level across 19 readability levels, with additional mappings to coarser 7, 5, and 3 level schemes. It supports multi-class readability classification tasks in 19, 7, 5, and 3 levels. Data instances include sentence text, word count, lexical information, readability levels (19, 7, 5, and 3 levels), annotator ID, document source, book, author, domain (Arts & Humanities, STEM, or Social Sciences), and text class (Foundational, Advanced, or Specialized). The dataset is split into train (80%), dev (10%), and test (10%) sets, balanced across readability levels, domains, and text classes.