Noor–Ghateh

Noor–Ghateh数据集是由伊朗科技大学和计算机研究中心伊斯兰科学（CRCIS）共同创建，专注于评估阿拉伯语在圣训领域的单词分割方法。该数据集包含约223,690个来自伊斯兰教法的单词，由专家标注，旨在通过其体积和多样性超越现有数据集。创建过程中，数据集通过多种方法如Farasa、Camel、Madamira和ALP进行评估，以确保标注质量。该数据集主要应用于阿拉伯语自然语言处理领域，特别是解决阿拉伯语单词分割问题，为研究人员提供了一个标准化的评估基准。

The Noor–Ghateh Dataset was jointly created by Iran University of Science and Technology and the Computer Research Center for Islamic Sciences (CRCIS). It is dedicated to evaluating word segmentation methods for Arabic in the Hadith domain. This dataset contains approximately 223,690 expert-annotated words sourced from Islamic jurisprudence, and is designed to outperform existing datasets in terms of both scale and diversity. During its development, the dataset was evaluated using multiple approaches including Farasa, Camel, Madamira and ALP to ensure annotation quality. Primarily utilized in Arabic natural language processing (NLP) research, this dataset specifically addresses Arabic word segmentation tasks, providing a standardized evaluation benchmark for researchers.