融合时序状态估计与非线性社区检测的社交网络回声室检测模型  

An Echo Chamber Detection Model for Social Networks Integrating Temporal State Estimation and Nonlinear Community Detection

作　　者：方尉旭 曹春萍[1] Weixu Fang;Chunping Cao(School of Optical-Electrical&Computer Engineering,University of Shanghai for Science&Technology,Shanghai)

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海

出　　处：《建模与仿真》2025年第3期652-666,共15页Modeling and Simulation

摘　　要：回声室检测工作是舆情防控中的基础任务。针对现有的回声室检测模型在考虑社区结构、节点属性、信息传播过程和时间因素时存在的不足,影响了节点间互动频率、信息传播概率、信息流动方向等特征的准确性的问题,本文提出了一种基于时序状态估计的社区回声室检测模型。利用改进的无迹卡尔曼滤波对社交网络进行状态估计,并设计融合节点极性和社区极性影响的传播模型。实验结果表明,与当前主流的滤波算法进行比较,该模型在社区中的回声室检测任务性能表现更好,证明所提出的模型能有效提升回声室检测的有效性,且在舆情管控中具有良好的应用价值。Echo chamber detection is fundamental in public opinion control.Addressing the limitations of existing models concerning the consideration of community structure,node attributes,information dissemination processes,and temporal factors,which affect the accuracy of features such as inter-activity frequency,information propagation probability,and information flow direction,this paper proposes a state estimation-based model for community echo chamber detection.Utilizing an im-proved Unscented Kalman Filter for state estimation of social networks and designing a propaga-tion model that incorporates node polarity and community polarity influences,the model outper-forms mainstream filtering algorithms in echo chamber detection tasks within communities.The experimental results demonstrate the effectiveness of the proposed model in enhancing echo cham-ber detection and its significant applicability in public opinion control.

关 键 词：社交网络 回声室 状态估计 信息传播 无迹卡尔曼滤波 社区结构 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]