CORE

该数据集是一个大规模的网络文本集合，这些文本根据其语域（即文体）进行了注释，包含了近5万篇文本，这些文本被归类为八大主要语域类别。该数据集采用两级分类法对语域进行编码，并展现了多种语言变体，这对分类任务提出了挑战。数据集规模接近5万篇文本，实验中共使用了38,760篇文档，任务类型为多标签分类。

This dataset is a large-scale web text corpus annotated based on register (i.e., writing style). It includes nearly 50,000 texts categorized into eight major register categories. The dataset adopts a two-level classification framework to encode registers and exhibits diverse language varieties, which poses challenges for classification tasks. A total of 38,760 documents from this corpus were employed in the corresponding experiments, and the task is multi-label classification.