社交媒体健康信息的语义分析：以推特上癌症相关推文为例  

Talking about Cancer on Twitter:Health Semantics and Social Media

作　　者：韩纲[1] 朱丹[2] 蔡承睿 王文[4] 

机构地区：[1]美国爱荷华州立大学Greenlee新闻与传播学院 [2]美国爱荷华州立大学商学院 [3]美国爱荷华州立大学工程学院 [4]美国爱荷华州立大学人类科学学院人类发展与家庭研究系

出　　处：《国际新闻界》2017年第4期44-62,共19页Chinese Journal of Journalism & Communication

摘　　要：本研究报告了对主流微博客社交媒体——推特(Twitter)16天内与癌症相关话题的语义分析。研究共收集了269万余条与癌症有关的推文(tweets),并创建了包含223条关键词的分类法(taxonomy)。依照推文的频率、周期、同步出现和情绪因素,分析了超过113万条由该分类筛选的推文并进行可视化呈现。研究结果发现:(1)可以从推特社交信息中检测到的、有助呈现癌症相关议题的最显见的关键词;(2)癌症相关推文的"每周两天"的频率高峰;这种节奏在很大程度上受到突发新闻或新闻事件的影响;(3)由与乳腺癌、肺癌和前列腺癌相关的推文中的关键词汇的同步呈现构成的语义网络(semantic network),以及(4)表达对癌症的积极或消极情绪的情感网络(sentiment network)。同时,本文对研究的潜在理论意义和实际应用进行了讨论。This study reports a semantic analysis of cancer-related conversation in Twitter during a 16- day period. More than 2.69 million tweets related to cancer were collected. Taxonomy consistingof 223 cancer-related key terms were created and developed. More than 1.13 million tweets filtered with the taxonomy were analyzed and visualized, in terms of the frequency, periodicity, and sentiments. Findings report （1） the most visible keywords, which partially illustrate the topics and message relevant to cancer, detectable from social streaming in Twitter; （2） a two-day-of-week rhythm with frequency of cancer-related tweets, which was highly influenced by breaking news or news events; （3） the key terms co-occurrence in tweets concerning breast cancer, lung cancer and prostate cancer, and （4） a sentiment network that comprises both positive and negative feelings or concerns about cancer. The potential theoretical contributions of this project and its practical implications are also discussed.

关 键 词：社交媒体 推特 健康信息 语义分析 大数据 

分 类 号：G206[文化科学—传播学] G254