THYME-2016

Clinical TempEval 语料库基于一组
600份癌症临床记录和病理报告
梅奥诊所的病人。这些笔记由 Mayo Clinic 手动去识别以替换
带有通用占位符的名称、位置等，
但时间表达没有改变。笔记
然后由 THYME 项目手动注释
(thyme.healthnlp.org) 使用 ISOTimeML 的扩展来注释临床笔记中的时间、事件和时间关系 (Styler et al., 2014b)。
此扩展包括新时间等附加功能
表达式类型（例如，用于表达式的 PREPOSTEXP
像术后），新的 EVENT 属性（例如，DEGREE=LITTLE 表示轻微恶心等表达），
并且更加关注类型的时间关系
包含（又名包括）。

The Clinical TempEval corpus is based on a set of 600 cancer clinical records and pathology reports from patients at the Mayo Clinic. These notes were manually de-identified by the Mayo Clinic, where names, locations and other personally identifiable information were replaced with generic placeholders, while temporal expressions remained unaltered. The notes were then manually annotated by the THYME project (thyme.healthnlp.org) using an extension of ISO-TimeML to annotate temporal information, events and temporal relations in clinical notes (Styler et al., 2014b). This extension includes additional features such as new temporal expression types (e.g., PREPOSTEXP for expressions like post-operative), new EVENT attributes (e.g., DEGREE=LITTLE for mild nausea and similar expressions), and greater emphasis on inclusion (also known as include) type temporal relations.

搜集汇总

背景与挑战

背景概述

THYME-2016是一个医疗领域的自然语言处理数据集，包含600份癌症临床记录和病理报告，用于时间信息提取、否定检测等任务。数据集由梅奥诊所手动去识别并注释，重点关注时间、事件和时间关系的标注。