基于机器学习算法和社会网络分析方法的议程设置验证——以新冠疫情初期的“中国援助”为例  

Verification on Agenda Setting Based on Machine Learning Algorithms and Social Network Analysis Methods—A Case Study of “Chinese Assistance” in the Early Stage of COVID-19

作　　者：万鑫 唐润发 熊令纯[1] 李裕梅[1] 

机构地区：[1]北京工商大学数学与统计学院,北京

出　　处：《数据挖掘》2022年第4期368-380,共13页Hans Journal of Data Mining

摘　　要：本文通过使用机器学习算法和社会网络分析方法,以新冠疫情初期的“中国援助”为例验证了议程设置的三个层次。首先,通过2020年3月1日~2020年4月30日的媒体报道量和百度指数,初步分析媒体报道量与公众关注程度的简单相关关系,得到两个结论:一是公众关于“中国援助”的关注度与媒体的报道量高度相关;二是当代媒介环境下,媒体影响公众的最佳时滞期由7天缩短到了当天。其次,采用支持向量机分类方法,将每篇媒体新闻报道与群众评论所蕴含的思想态度分成积极正向和消极负向两方面,进而分析媒体新闻报道对群众评论的主观思想的影响,通过分析认为媒体对疫情下“中国援助”偏向积极地报道带动了公众偏向积极地看待中国援助国外。最后, 通过TextRank算法和社会网络分析方法,构建媒体属性矩阵和公众属性矩阵,并分析媒体报道和公众评论的相关性,发现媒体的报道推动了公众对“中国援助”事件的认知网络的构建。Machine learning algorithms and social network Analysis methods are used to verify the three levels of agenda setting in the early stage of “Chinese assistance” in COVID-19. Firstly, through the media coverage and Baidu Index from March 1, 2020 to April 30, 2020, this paper preliminarily analyzes the simple correlation between media coverage and public attention, and two conclusions are obtained: the public’s attention to “Chinese assistance” is highly related to the amount of media coverage, thus the first level of agenda setting is verified;in the contemporary media environment, the best time lag period for media to affect the public has been shortened from 7 days to the same day. Secondly, by using SVM model, the thoughts and attitudes contained in each media news report and public comments are divided into positive and negative aspects, and then the impact of media coverage on the subjective thoughts of public comments is analyzed. Through the analysis, it is considered that the positive media coverage of “Chinese assistance” has led the public to view Chinese assistance to foreign countries positively, thus the second level of agenda setting is verified. Finally, through TextRank algorithm and social network analysis methods, the media attribute matrix and public attribute matrix are constructed, and the correlation between media coverage and public comments is analyzed. It is concluded that media coverage affects the public’s cognitive network construction about the “Chinese assistance” event, thus the third level of agenda setting is verified.

关 键 词：议程设置 新闻报道 属性矩阵 机器学习算法 媒体报道 百度指数 媒介环境 社会网络分析方法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] R181.8[自动化与计算机技术—控制科学与工程]