PHM2017

PHM2017 是一个新数据集，包含 7,192 条英文推文，涵盖六种疾病和状况：阿尔茨海默病、心脏病（任何严重程度）、帕金森病、癌症（任何类型）、抑郁症（任何严重程度）和中风。 Twitter 搜索 API 用于使用口语疾病名称作为搜索关键字来检索数据，期望检索高召回率、低精度的数据集。删除转发推文和回复后，推文被手动注释。标签是：自我提及。该推文包含一个健康提及以及 Twitter 帐户所有者的健康自我报告，例如，“但是，我努力工作，并在 2014 年 1 月至 2 月期间参加了东京市长选举活动，但没有公开癌症。”其他提及。该推文包含有关帐户所有者以外的其他人的健康报告的健康提及，例如，“患有帕金森氏症的设计师无法工作，然后工程师发明了手镯 + 改变了她的世界”意识。推文包含疾病名称，但没有提及具体的人，例如，“心脏病发作前一个月，你的身体会用这 8 个信号警告你”非健康。推文包含疾病名称，但推文主题与健康无关。 “现在我可以把癌症挂在墙上，让所有人都能看到 <3”

PHM2017 is a novel dataset consisting of 7,192 English tweets covering six diseases and health conditions: Alzheimer’s disease, heart disease (of any severity), Parkinson’s disease, cancer (of any type), depression (of any severity), and stroke. The Twitter Search API was used to retrieve the data with colloquial disease names as search keywords, with the expectation of obtaining a dataset with high recall and low precision. After removing retweets and replies, the tweets were manually annotated. The annotation labels and their definitions are as follows: 1. Self-mention: Tweets containing a health mention and a health self-report from the Twitter account owner, e.g., "However, I worked hard and campaigned for the Tokyo mayoral election between January and February 2014, but did not publicly disclose my cancer." 2. Other-mention: Tweets containing health mentions reporting on the health of people other than the account owner, e.g., "A designer with Parkinson’s disease could no longer work, then an engineer invented a bracelet + changed her world" 3. Awareness: Tweets containing disease names but without mentioning specific individuals, e.g., "One month before a heart attack, your body will warn you with these 8 signs" 4. Non-health: Tweets containing disease names but whose topic is unrelated to health, e.g., "Now I can hang my cancer on the wall for everyone to see <3"

搜集汇总

背景与挑战

背景概述

PHM2017是一个包含7,192条英文推文的数据集，涵盖六种疾病和状况，推文经过手动注释分为四类。该数据集由埃默里大学于2017年发布，适用于医疗数据分析和流行病学研究。