信息资源管理学报 ›› 2021, Vol. 11 ›› Issue (3): 121-131.doi: 10.13365/j.jirm.2021.03.121

• 研究论文 • 上一篇    下一篇

基于多维特征的网络健康信息可信度评估研究

陆泉1,2 岳雪琪1 刘婷1 李易时1 陈静3   

  1. 1.武汉大学信息资源研究中心; 
    2.武汉大学大数据研究院,武汉,430072; 
    3.华中师范大学信息管理学院,武汉,430079
  • 出版日期:2021-05-26 发布日期:2021-06-23
  • 作者简介:陆泉,教授,博士,研究方向为数据挖掘、人机交互、决策支持系统、知识组织与知识服务等;岳雪琪,硕士生,研究方向为数据挖掘、知识服务;刘婷,博士生,研究方向为用户行为、知识服务;李易时,硕士生,研究方向为数据挖掘、知识组织;陈静(通讯作者),教授,博士,研究方向为知识管理、数字图书馆、数据挖掘等,Email:jchen@mail.ccnu.edu.cn。
  • 基金资助:
    本文系国家社会科学基金重点项目“心理账户理论视角下在线健康社区精准信息服务研究”(20ATQ008)的研究成果之一。

Research on the Credibility Assessment of Network Health Information Based on Multidimensional Information Characteristics

Lu Quan1,2 Yue Xueqi1 Liu Ting1 Li Yishi1 Chen Jing3   

  1. 1. Center for Studies of Information Resources at Wuhan University; 
    2. Big Data Research Institute of Wuhan University, Wuhan, 430072; 
    3. School of Information Science at Central China Normal University, Wuhan,430079
  • Online:2021-05-26 Published:2021-06-23

摘要: 网络健康信息可信度评估是一个比较复杂的新问题,但可以根据现有理论从多维信息特征展开思考。本研究旨在建立一套网络健康信息的可信度评估指标体系,可用于搜索引擎的网络健康信息评估过滤,也可为用户提供可信度评估的指导建议。从信息可信度的内涵出发,提出了信息内容、信息源以及传播媒介三个维度下的多个信息特征,设计了网络健康信息可信度评估指标体系,并通过层次分析法确定指标权重,利用自然语言处理技术和赋值转化的方法实现各指标的自动测度。通过实证研究验证评估指标体系的有效性,选取了疾病预防、疾病症状和疾病治疗这三类典型的网络健康信息搜索主题进行实证。将指标体系评估得分与专家标注的标准集进行对照分析发现,该评估指标体系可有效评估网络健康信息可信度,在可信度二分类下准确率与精确率达到90%以上,而在可信度五分类时评估得分阈值区间呈现为两头宽中间窄,说明可高效区分可信与不可信网络健康信息。

关键词: 信息可信度, 网络健康信息, 信息特征, 评价指标

Abstract: The credibility assessment of network health information is a complex problem proposed recently, which can be considered from the perspective of multidimensional information characteristics according to the existing theories. This paper aims to establish a credibility evaluation index system of network health information, which can be used to evaluate and filter the network health information for search engines, and also provide guidance and suggestions of credibility evaluation to information users.According to the connotation of information credibility, this paper puts forward three dimensions of information content, information source and communication media, designs the evaluation index system of network health information credibility, determines the index weight through analytic hierarchy process, and uses natural language processing technology and assignment transformation to measure indexes automatically. Disease prevention, disease symptoms and disease treatment, these three typical themes in the context of Internet health information search are considered in empirical research. Comparing the system score with the ‘Gold Standard’ labeled by the two experts, the credibility evaluation index system is proved effectively to evaluate the credibility of network health information. Under the credibility dichotomy, the evaluation accuracy and precision ratio reached more than 90%, and the scoring threshold interval of credibility under five classifications is wide at both ends and narrow in the middle, which could effectively distinguish the trusted and untrusted network health information.

Key words: Information credibility, Network health information, Information characteristics, Assessment index

中图分类号: