一种干扰系数自探测的网络事件选取方法  

Network Event Selection Method Based on Interference Coefficient Self-detection

作　　者：臧洁[1] 任旭 冯艳爽 王妍[1] 肖萍[3] 鲁锦涛 ZANG Jie;REN Xu;FENG Yanshuang;WANG Yan;XIAO Ping;LU Jintao(Liaoning University,College of Information,Shenyang 110036,China;Rongke Technology Co.,Ltd.Shenyang 110027,China;Criminal Investigation Police University of China,Public Security Information Technology and Information Institute,Shenyang 110854,China)

机构地区：[1]辽宁大学信息学院,沈阳110036 [2]荣科科技股份有限公司,沈阳110027 [3]中国刑事警察学院公安信息技术与情报学院,沈阳110854

出　　处：《小型微型计算机系统》2024年第3期763-768,共6页Journal of Chinese Computer Systems

基　　金：国家重点研发计划项目(2019YFB1405804)资助.

摘　　要：网络事件在传播过程中会产生相互影响,对这种事件间的影响进行研究有助于深入了解网络舆情的传播规律,并依此采取正确的应对办法.针对不同事件间存在不同干扰效果这一问题,提出了一种干扰系数自探测的网络事件选取方法.该方法使用具有多事件容量特征的mSIR-CA(MultipleSIR-Cellular Automata)模型为推演核心,并借助真实事件数据进行有针对性的参数寻优,使得方法能够根据目标事件,从当前同处于扩散过程的多个备选干扰事件中,进行有效事件选取及干扰效果排序.实验结果表明,根据真实事件训练集对方法参数寻优后,通过对已发生事件标记的事件组进行验证,该方法在有效干扰事件的分类判断、干扰效果排序结果方面有较好的准确率.Network events affect each other in the process of communication,the research on the influence of this kind of event is helpful to deeply understand the propagation law of network public opinion,and take the correct response accordingly.Aiming at the problem that different events have different interference effects,a network event selection method based on interference coefficient self-detection was proposed.In this method,the mSIR-CA(Multiple SIR-Cellular Automata)model with multi-event capacity characteristics was used as the deduction core,and the real event data was used for targeted parameter optimization.According to a target event,the method could carry out effective event selection and interference effect sequence from a plurality of alternative interference events which were in the diffusion process at present.After the parameters are optimized by according to the training set of real events and the event groups marked by events are verified,the experimental results show that the method has good accuracy in the classification and judgement of effective interference events,and the sorting results of interference effects.

关 键 词：网络舆情 mSIR-CA模型 干扰事件选取 参数寻优 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]