Learning health information from user-generated Web content

从用户生成的 Web 内容中学习健康信息

• 批准号: 402096-2011
• 负责人: Sokolova, Marina
• 金额: $ 1.02万
• 依托单位: University of Ottawa
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2015
• 资助国家: 加拿大
• 起止时间: 2015-01-01 至 2016-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=573158
• 关键词: Learning; health; information; user; generated

Social networks, virtual communities, blogs became a major means of human communication. Information posted in the user-generated Web content has promoted development of a new class of text data mining methods. Those methods aim to find a new knowledge from non-contrived text of the user-generated content. I was one of the first investigators to develop text mining methods of user-generated texts. My proposed research program will build on this earlier body of work. The proposed research program is to develop methods for extracting health information from user-generated content found in social media; including details pertaining to personal or public health (e.g., ailment, symptoms). The methods will employ natural language processing, information extraction and machine learning approaches. These new methods will serve two important purposes; they will allow us to determine whether the new modes of communication are leading to an increase in privacy breaches and they will facilitate the mining of these new forums for new health knowledge. One specific goal of this program is the creation of a repository of keywords and phrases that would be made available to other investigators working in the fields of data mining, language technologies and health research. Unsolicited disclosure of confidential health information can result in legal and security problems therefore early identification of possible confidentiality breaches is critical. Conversely, detected health information can contribute to public well-being; for example, reports of same symptoms from the same geographic area could signal a starting stage of an epidemic.

社交网络、虚拟社区、博客成为人类交流的主要手段。在用户生成的Web内容中发布的信息促进了一类新的文本数据挖掘方法的发展。这些方法旨在从用户生成内容的非人工文本中发现新的知识。 我是最早开发用户生成文本的文本挖掘方法的研究者之一。我提出的研究计划将建立在这个早期的工作机构。 拟议的研究计划是开发从社交媒体中发现的用户生成内容中提取健康信息的方法;包括与个人或公共健康有关的细节（例如，疾病、症状）。 这些方法将采用自然语言处理、信息提取和机器学习方法。 这些新方法将服务于两个重要目的;它们将使我们能够确定新的通信模式是否导致隐私泄露的增加，它们将促进挖掘这些新的健康知识的新论坛。 该计划的一个具体目标是创建一个关键字和短语库，供在数据挖掘、语言技术和健康研究领域工作的其他调查人员使用。 未经请求披露机密健康信息可能会导致法律的和安全问题，因此及早识别可能的机密性违规至关重要。 相反，检测到的健康信息可以促进公众福祉;例如，来自同一地理区域的相同症状的报告可能标志着流行病的开始阶段。