Since the advent of online social media platforms such as Twitter and Facebook, useful health-related studies have been conducted using the information posted by online participants. Personal health-related issues such as mental health, self-harm and depression have been studied because users often share their stories on such platforms. Online users resort to sharing because the empathy and support from online communities are crucial in helping the affected individuals. A preliminary analysis shows how contents related to non-suicidal self-injury (NSSI) proliferate on Twitter. Thus, we use Twitter to collect relevant data, analyse, and proffer ways of supporting users prone to NSSI behaviour. Our approach utilises a custom crawler to retrieve relevant tweets from self-reporting users and relevant organisations interested in combating self-harm. Through textual analysis, we identify six major categories of self-harming users consisting of inflicted, anti-self-harm, support seekers, recovered, pro-self-harm and at risk. The inflicted category dominates the collection. From an engagement perspective, we show how online users respond to the information posted by self-harm support organisations on Twitter. By noting the most engaged organisations, we apply a useful technique to uncover the organisations' strategy. The online participants show a strong inclination towards online posts associated with mental health related attributes. Our study is based on the premise that social media can be used as a tool to support proactive measures to ease the negative impact of self-harm. Consequently, we proffer ways to prevent potential users from engaging in self-harm and support affected users through a set of recommendations. To support further research, the dataset will be made available for interested researchers.
Detecting potentially harmful and protective suicide-related content on twitter: A machine learning approach
Medical University of Vienna(维也纳医科大学);Complexity Science Hub Vienna(维也纳复杂性科学中心);Graz University of Technology(格拉茨工业大学);Institute of Globally Distributed Open Research and Education(全球分布式开放研究与教育研究所);Vienna University of Technology(维也纳工业大学)
An Exploratory Study of Tweets about the SARS-CoV-2 Omicron Variant: Insights from Sentiment Analysis, Language Interpretation, Source Tracking, Type Classification, and Embedded URL Detection
University of Cincinnati(辛辛那提大学)
本文介绍了一项探索性研究的结果,该研究针对Twitter上不断生成的大数据,这些数据与分享有关COVID-19大流行的信息、新闻、观点、意见、想法、反馈和经验有关,特别关注Omicron变体,这是目前SARS-CoV-2的全球主要变体。共研究了12028条关于Omicron变体的推文,并分析了推文的具体特征,包括情感,语言,来源,类型和嵌入式URL。这项研究的结果是多方面的。首先,从情感分析中,观察到50.5%的推文具有中性情感。其他情绪-坏的,好的,可怕的和伟大的分别在15.6%,14.0%,12.5%和7.5%的推文中被发现。其次,语言翻译的结果显示,65.9%的推文是用英语发布的。其次是西班牙语、法语、意大利语和其他语言。第三,来源追踪的结果显示,Android版Twitter与35.2%的推文相关。紧随其后的是Twitter Web App、iPhone版Twitter、iPad版Twitter和其他来源。第四,研究推文的类型发现,转发占推文的60.8%,其次是原创推文和回复,分别占推文的19.8%和19.4%。第五,在嵌入URL分析方面,发现最常见的嵌入推文的域名是twitter.com,其次是biorxiv.org,nature.com和其他域名。最后,为了支持该领域的类似研究,我们开发了一个Twitter数据集,该数据集包含自2021年11月24日首次发现该变体病例以…