GRDD: A Dataset for Greek Dialectal NLP
收藏资源简介:
GRDD数据集由克里特大学创建,旨在为现代希腊方言的自然语言处理提供资源。该数据集包含来自四种现代希腊方言(克里特、庞蒂克、北部希腊和塞浦路斯希腊)的原始文本数据,总计约230万字,尽管存在不平衡。数据集的创建过程涉及从互联网上自由获取的方言数据,包括博客、网站和文学文本,并进行了基本的预处理。该数据集主要用于方言识别任务,展示了即使是简单的机器学习模型也能有效区分这些方言。
The GRDD dataset was created by the University of Crete to provide resources for natural language processing (NLP) of modern Greek dialects. It contains raw textual data from four modern Greek dialects: Cretan, Pontic, Northern Greek, and Cypriot Greek, totaling approximately 2.3 million words, though it is imbalanced. The dataset construction involved freely acquired dialectal data from the Internet, including blogs, websites, and literary texts, with basic preprocessing performed. This dataset is primarily used for dialect identification tasks, demonstrating that even simple machine learning models can effectively distinguish between these dialects.




