摘要
[目的/意义]基于意见领袖专业权威成长动态性,探讨意见领袖预测指标,在网络舆情爆发初期快速、准确预测意见领袖,合理设置正面舆论议程、引导舆论。[方法/过程]选择年度影响力较大的10个网络舆情事件的用户信息、历史文本数据,通过话题提取等方法构建用户活跃度、历史影响力、专业权威等共3类15个指标,构建预测模型,分析不同指标体系的预测精度差异、指标影响力差异,并提取预测精度最高的指标集合。[结果/结论]基于临近事件期的历史文本提取的话题相似度指标、转评赞数据构建的历史影响力指标,较为细致地刻画了意见领袖专业权威成长轨迹,对特定话题的意见领袖预测具有重要意义。
[Purpose/Significance]Based on the dynamic growth of E-opinion leaders’ professional authority, this paper discusses the prediction indicators of opinion leaders, which may provide suggestions for government department to quickly and accurately predict E-opinion leaders, reasonably set positive public opinion agenda and guide public opinion in the early stage of the outbreak of Internet public opinion. [Method/Process] We collect the user’s information and historical text data based on 10 internet public opinion events, and construct 15 predict indicators by topic extraction, including user’s activity, historical influence and professional authority. Then we construct prediction models to analyze the difference of prediction accuracy and influence of these 15 indicators. And we extract the index set which has the lowest accuracy error. [Result/Conclusion] The topic similarity indicators extracted from the historical text, and the historical influence indicators calculated by the number of forwarded, agreed, commented, describe the growth track of e-opinion leaders’ professional authority in detail, which have great significance to the prediction of e-opinion leaders on specific topics.
作者
林萍
王晓梅
吕健超
黄卫东
Lin Ping;Wang Xiaomei;Lyu Jianchao;Huang Weidong(School of Management,Nanjing University of Posts and Telecommunications,Nanjing 210003;Research Base of Philosophy and Social Sciences in Jiangsu--Center of Innovation and Emergency Management in Information Industry,Nanjing 210003;Jiangsu Joint Postgraduate Training Base Construction by Nanjing University of Posts and Telecommunicationand Socool—Tech Co.,Ltd,Nanjing 210000)
出处
《情报杂志》
CSSCI
北大核心
2021年第10期59-65,共7页
Journal of Intelligence
基金
江苏省高校优秀中青年教师和校长境外研修计划项目成果之一
国家自然科学基金面上项目“基于个体情感模型的舆情传播机制研究”(编号:71671093)
江苏省学术学位研究生创新计划项目“基于信息质量视角的网络舆情传播关键节点识别研究”(编号:KYCX19_0992)。
关键词
意见领袖
专业权威
成长动态
意见领袖预测
e-opinion leader
professional authority
growth track
prediction of e-opinion leaders