i2b2-2012

该数据集名为i2b2-2012挑战数据集，专为时间关系抽取而设计，包含了临床文本以及事件间额外的实体类别和时间关系。该数据集分为70%的训练集、15%的验证集和15%的测试集，其格式与i2b2-2009不同，并增加了额外的类别。规模上，该数据集包含760个文件，其中包括190个TXT文件和190个XML文件用于标注。任务方面，主要是进行时间关系抽取。

This dataset, named the i2b2-2012 Challenge Dataset, is specifically designed for temporal relation extraction. It contains clinical texts, as well as additional entity categories and temporal relations between events. The dataset is split into 70% training set, 15% validation set and 15% test set. Its format differs from that of the i2b2-2009 dataset, with additional categories added. In terms of scale, the dataset includes 760 files in total, including 190 TXT files and 190 XML annotation files. The main task of this dataset is temporal relation extraction.