摘要
基于双层异步社会网络研究口碑病毒式传播的影响最大化问题,通过扩展经典的独立级联模型,构建相应的影响最大化模型,设计多种基于启发式算法的病毒式营销策略选取种子节点。针对所提算法,从网络传播概率、网络结构和传播异步性等3个方面进行仿真对比研究。结果表明:该设计方法明显优于传统的随机规则和出度规则,但不同方法适用于不同的传播概率和网络结构;随着双层网络传播异步性的增加,各算法营销结果均出现明显下降。
The influence maximization problem based on viral diffusion of word-of-mouth in the context of bi-level asynchronous social networks was studied.Specifically,the corresponding influence maximization model was proposed by extending the classical interdependent cascade model.Based on heuristic algorithms,several viral marketing strategies were proposed to select seed nodes.Simulations based on proposed strategies with different diffusion probabilities,network structures and diffusion asynchronous extents were implemented respectively.Results show that the proposed methods outperform classical random rule and max-degree rule,and they are suitable to different settings of diffusion probabilities and diffusion asynchronous extents.The marketing results of all these methods will deteriorate with the increase of diffusion asynchronous extents.
作者
王长军
王葛格
WANG Changjun;WANG Gege(Glorious Sun School of Business and Management,Donghua University,Shanghai 200051,China;Hangzhou Best Network Technologies Co.Ltd.,Hangzhou 310013,China)
出处
《东华大学学报（自然科学版）》
CAS
北大核心
2021年第4期107-115,共9页
Journal of Donghua University(Natural Science)
基金
国家自然科学基金重点资助项目(71832001)
国家自然科学基金资助项目(71872036)
上海市哲学社会科学规划基金资助项目(2019BGL036)
上海市自然科学基金资助项目(20ZR1401900)
中央高校基本科研业务费专项资金服务管理与创新基地资助项目(2232018H-07)。
关键词
病毒式营销
营销算法
影响最大化
双层网络
独立级联模型
viral marketing
marketing algorithm
influence maximization
bi-level network
independent cascade model
