摘要
[目的/意义]针对前人成果普遍存在的处理效率与识别准确度等方面的问题,提出了基于复杂网络理论的微博谣言识别与预警算法。[方法/过程]该算法在第一阶段基于复杂网络理论,以动态节点管理为基础,对微博节点实施行为刻画,实现谣言恶意散布节点的早期告警与预测;在第二阶段以复杂网络关系发展理论为基础,将谣言与真实信息进行信息轨迹聚类与隔离,最终实现谣言侦测与识别。'[结果/结论]基于新浪与腾讯微博作为数据源的实验证明,该算法较之既往算法,不但谣言覆盖度与识别准确率高,而且响应速度快、处理效率高,具有良好的性价比。
[Purpose/Significance]Based on the common processing efficiency and recognition accuracy of predecessors’ results, a microblog rumor recognition and early warning algorithm based on complex network theory is proposed. [Method/Process]Based on the complex network theory in the first stage, the algorithm is based on dynamic node management, and describes the behavior of microblog nodes to realize the early warning and prediction of rumor malicious distribution nodes. In the second stage, based on the theory of complex network relationship development, the trajectory is clustered and isolated from the trajectory of the rumor and the real information, and finally the rumor detection and recognition is realized.[Results/Conclusion]Experiment results based on Sina and Tencent micro-blog as data sources show that compared with the previous algorithm, this algorithm not only has high accuracy in coverage and recognition of gossip, but also has fast response speed, high processing efficiency and good cost performance.
作者
王征
叶长安
Wang Zheng;Ye Changan(Economic Information Engineering Southwestern University of Finance and Economics,Chengdu 611130;Marxism School,Southwestern University of Finance and Economics,Chengdu 611130)
出处
《情报杂志》
CSSCI
北大核心
2019年第4期148-154,共7页
Journal of Intelligence
基金
中央高校基本科研业务费项目"基于群体自动叙事的互联网金融市场情绪预测模型研究"(编号:JBK1903006)
中央高校教育教学改革专项-研究生教育教学改革项目"互联网舆情管理与分析"(编号:JYJ20180204)成果之一
关键词
舆情监测
谣言识别
两阶段法
行为识别
复杂网络
早期预警
public opinion gossip recognition
two-stage method
behavior recognition
complex network early warning
