摘要
从用户影响力和用户活跃度两个角度考虑,构建了微博意见领袖指标体系,提出了使用层次分析法和粗糙集决策分析理论对意见领袖的特征进行识别及分析的理论框架。对3起突发事件微博数据进行了实证分析,通过观察法分析了Top 10意见领袖具有的特征,使用粗糙集理论对意见领袖识别问题进行了建模,并提取了识别意见领袖的决策规则。另外,通过定义意见领袖一致性和排名稳定性两个指标分析了意见领袖的跨主题性,实验结果表明意见领袖是主题依赖的,只有很少用户可以在不同主题同时成为意见领袖。
This paper constructs an opinion leader index system using the user influence and user activeness,and analyzes the weights of indexes with AHP（Analytic Hierarchy Process）.This research analyzes the characteristics of public opinion based on the dataset from 3 emergency events in sina microblog, analyzes the characteristics of top 10 opinion leaders,and extracts the decision rules based on the rough set theory.We define two indicators： the consistency of opinion leaders and the stability of rank,and study the opinion leader＇s cross-topic characteristic.The experimental results demonstrate that the opinion leaders are topic-dependent and that only a few users can become opinion leaders across different topics.
出处
《系统工程》
CSSCI
CSCD
北大核心
2011年第6期8-16,共9页
Systems Engineering
基金
国家自然科学基金资助项目(90924020)
高等学校博士学科点专项科研基金资助项目(200800060005)
关键词
突发事件
意见领袖
网络舆情
微博
用户影响力
用户活跃度
Emergency Event
Opinion Leader
Public Opinion
Microblog
User Influence
User Activeness
