Cantemist corpus: gold standard of oncology clinical cases annotated with CIE-O 3 terminology

Cantemist shared task train and development sets with Gold Standard annotations AND Cantemist test+background set without annotations. It contains the train and development sets of the three subtasks: cantemist-ner, cantemist-norm and cantemist-coding. In addition, it contains the documents of the test and background sets, without annotations. For subtasks cantemist-norm and cantemist-ner, annotations are distributed in Brat format. See Brat webpage for more information For subtask cantemist-coding, codes are grouped in a TSV file with the following columns (this follows the format used in CodiEsp shared task): filename code In the three subtasks, the goal will be to predict the annotations (either the ANN files or the TSV with the codes) given only the plain text files. In addition, it contains the plain text files of the test and background sets (at this point, they are released together). For further information, please visit https://temu.bsc.es/cantemist/ or email us at [e-mail]本数据集包含Cantemist共享任务的训练集与开发集（带有金标准标注（Gold Standard annotations）），以及无标注的Cantemist测试集与背景集。其涵盖三个子任务：cantemist-ner、cantemist-norm与cantemist-coding的训练集与开发集；此外还包含无标注的测试集与背景集文档。针对cantemist-norm与cantemist-ner子任务，标注采用Brat格式存储，详细信息可查阅Brat官方网页。针对cantemist-coding子任务，相关代码收纳于TSV文件中，该文件遵循CodiEsp共享任务的格式规范，包含两列内容：文件名（filename）与代码（code）。三个子任务的核心目标均为：仅基于纯文本文件，预测对应标注内容（即ANN标注文件或包含代码的TSV文件）。此外，本次发布同步提供测试集与背景集的纯文本文件。如需获取更多信息，请访问https://temu.bsc.es/cantemist/ 或发送邮件至[e-mail]