摘要
采用传染病模型对网络热点话题的传播进行建模具有重要的价值,但是现有的传染病模型并没有区分话题类型和不同用户传播话题的概率,为此提出一个基于内容的网络热点话题传播模型.模型中引入了用户对话题传播的敏感度,基于用户话题敏感度定义了单个用户传播话题的概率,融合话题的内容分类特性、用户传播概率、用户重入概率等因素,借鉴SIRS模型的基本思想,构建了话题传播模型(CSIRS).在无标度网络、小世界网络、随机网络和真实社会网络上作了不同实验,实验结果表明CSIRS模型不仅能够呈现一般传染病动力模型的传播模式,还能够呈现多个波动、小范围长时间传播、快速上升缓慢下降等社会网络热点话题的传播模式.该模型为融合网络结构和话题内容属性建模话题传播过程带来新的研究思路.
Current researches suggest a new hot topic propagation modeling pattern, which uses the epidemic mod- el, and adds great value to the industry. Unfortunately, the existing epidemic models make no distinction between topic type and topic propagation probability of different users. A new propagation model based on topic contents for network hot topics was proposed in this paper. User＇s sensitivity degree to topic propagation was introduced and based on this, the single user propagation probability was defined. By integrating factors such as topic contents clas- sification characteristics, user propagation probability, user re-entry probability, and drawing on the basic idea of SIRS model, a topic propagation model （CSIRS for short） was built. Different experiments were conducted respec- tively in a scale-free network, small-world network, random network and real social network. The experimental re- suits show CSIRS model can not only present the propagation pattern of a general dynamic model of infectious dis- ease, but also present the propagation patterns of hot topics on social networks, such as multiple wave propagation patterns, small scale spreading with long tail propagation pattern and rapidly rising slowly falling tern. CSIRS model provides a new idea for modeling the topic propagation process by integrating and topic attribute.
出处
《智能系统学报》
CSCD
北大核心
2013年第3期233-239,共7页
CAAI Transactions on Intelligent Systems
基金
国家自然科学基金资助项目(61170112)
教育部人文社会科学研究基金资助项目(13YJC860006)
北京市属高等学校科学技术与研究生教育创新工程建设项目(PXM2012_014213_000037)
关键词
热点话题
传播模型
传染病模型
话题传播模型
hot topic
propagation model
epidemic model
topic propagation model propagation pat- network topology
