首页 | 本学科首页   官方微博 | 高级检索  
     

微博“树洞”留言的负性情绪特征分析
引用本文:陈盼,钱宇星,黄智生,赵超,刘忠纯,杨冰香,杨芳,张晓丽. 微博“树洞”留言的负性情绪特征分析[J]. 中国心理卫生杂志, 2020, 0(5): 437-444
作者姓名:陈盼  钱宇星  黄智生  赵超  刘忠纯  杨冰香  杨芳  张晓丽
作者单位:武汉大学健康学院;武汉大学信息资源研究中心;荷兰阿姆斯特丹大学人工智能系;华中农业大学信息学院;武汉大学人民医院精神卫生中心
基金项目:教育部人文社会科学研究青年基金项目(20YJCZH204);国家重点研发计划(2018YFC1314600);武汉大学国家大学生创新创业训练计划项目(201910486111)。
摘    要:目的:探讨微博"树洞"留言的情绪特征,为开展心理健康促进、自杀预警及干预提供思路和数据支持。方法:通过"树洞"智能机器人抓取留言数据351 825条,通过分词、TF-IDF算法得到词频排名前500的高频关键词,利用Gephi软件对关键词进行共现网络分析,利用Boson提供的情感词典判断所提取高频关键词的情感正负程度,计算每个小时内留言的数量,并对负性情绪进行内容分析。结果:发布留言在22∶00~02∶00活跃度最高(占总留言量的36.3%),05∶00~07∶00活跃度最低(占总留言量的4.2%)。数据经处理后的346 075条有效留言数据中,偏正性情感表达的有165 890条,其中情感得分大于0的留言有137 132条,占总留言量的39.7%,等于0的有28 578条(8.2%),偏负性情感表达的留言有180 185条,占总留言量的52%。各时间段负性情绪的表达均多于正性情绪。留言的内容包括情绪的倾诉、人际关系和社会支持、睡眠以及死亡等方面,在表达自杀意念的人群中,"跳楼、安眠药、割腕、烧炭"是提及自杀方式最多的词。结论:微博"树洞"用户普遍表达的负性情绪以及相关负性生活事件应引起人...

关 键 词:树洞微博  人工智能  网络爬虫  情感分析  社交媒体

Negative emotional characteristics of Weibo"Tree Hole"users
CHEN Pan,QIAN Yuxing,HUANG Zhisheng,ZHAO Chao,LIU Zhongchun,YANG Bingxiang,YANG Fang,ZHANG Xiaoli. Negative emotional characteristics of Weibo"Tree Hole"users[J]. Chinese Mental Health Journal, 2020, 0(5): 437-444
Authors:CHEN Pan  QIAN Yuxing  HUANG Zhisheng  ZHAO Chao  LIU Zhongchun  YANG Bingxiang  YANG Fang  ZHANG Xiaoli
Affiliation:(School of Health Sciences,Wuhan University,Wuhan 430071,China;Center for the Studies of Information Resources,Wuhan University,Wuhan 430072,China;Division of Mathematics and Computer Science,Faculty of Sciences,Vrije University Amsterdam,Amsterdam 1081hv,Netherland;School of Information Sciences,Huazhong Agricultural University,Wuhan 430070,China;Department of Psychiatry,Renmin Hospital of Wuhan University,Wuhan 430060,China)
Abstract:Objective:To explore the negative emotional characteristics of Weibo"Tree Hole"messages in order to provide ideas and data support for mental health promotion,suicide waning and intervention.Method:U­sing a"Tree Hole"artificial intelligence(AI)robot,351825 messages were reviewed and words were divided by using a Jieba tool.The 500 most frequently used key words were calculated using a TF-IDF algorithm.Keywords co-occurrence network was constructed for social network analysis with important keywords by using Gephi soft­ware.Boson affective dictionary was used to judge for emotion characteristics of high frequency keywords.The a­mount of messages per hour was collected and the content characteristics of negative emotionswere analyzed.Results:Most messages were posted between 22:00 and 02:00,accounting for 36.3%of the total number of messages reviewed.The time period for posting the fewest number of messages was 05:00 to 07:00,accounting for 4.2%of the total number of messages reviewed.Among the 346075 valid messages after data processing,165890 messages were positive emotional expression,among which 137132 messages had an emotional score grea­ter than 0,accounting for 39.7%of the total messages,28578 messages had an emotional score equal to 0(8.2%),and 10185 messages were negative emotional expression,accounting for 52%of the total messa­ges.Negative emotions were more expressed than positive emotions in each time period.The contents of the mes­sage included emotional expression,interpersonal relationship and social support,sleep and death.Among the people who express suicidal ideation,jumping off the building,sleeping pills,cutting wrists and burning charcoal were the most mentioned words related to suicidal ways.Conclusion:Negative emotions and related negative life events commonly are expressed by users of Weibo"tree hole".More attention should be paid to this phenomena,especially for users who express suicidal ideation,early identification and interventions are important.
Keywords:Tree hole of Weibo  artificial intelligence  Web Crawlers  affective analysis  social media
本文献已被 CNKI 维普 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号