Construction Of Health Knowledge Graph Based On Multi-source UGC Data

The concept of consumers’health care has changed from getting treatment passively to managing and sharing health issues positively. And The past medical model of consumers was going to hospital, which has also been moved to focusing on prevention, daily health care and rehabilitation.Therefore, academia area, medicine area and business area are actively exploring how to provide consumers with more accurate, convenient and comprehensive medical and health services. And the basement of the accurate medical services is to identify the health topic. However, in the context of the current study, the study of consumer health topic analysis usually focus on one type of specific diseases, using the data samples directly from the consumers, and their research methods are mostly used questionnaire survey method, experimental method,or interview method. These are likely to lead to the conclusion is not comprehensive, and the applicability is not strong. Thence, this article use the methods combined with qualitative and quantitative, applying to the data, which is coming from social media with various diseases, and then refine the health topics, extract the feature words, finally put forward the consumer health knowledge map construction, and discusses the revelation for improving consumer health information literacy and health information system design.Besides, this article discusses the three applications of knowledge map.Innovation of this article is mainly in the following:(1) the innovation of the research object. In this paper, a combination of four types of social media platforms in three different diseases of UGC data as the research object, try to maximum cover different text types of disease of health information text, on the basis of constructing the consumer focus on health and subject system, and draw maps and related conclusions.(2) the innovation of research methods application. This paper puts forward the method of qualitative method-open coding combined with quantitative research methods text mining, completing theme and feature words refining. It helps extracting multi-level characteristics effectively from a mass data sample, also helps with data visualization. And the graphic database is used as a solution of knowledge graph drawing.(3) The innovation of the research results. In this paper, through research and analysis, we have found some interesting conclusions:If consumer attention to all kinds of disease are consistent, consumers in all kinds of diseases of different themes attention exist obvious difference; There are distance distinguishes between different health theme; Words from different social media language in describing the same kind of disease is different.