电子商务系统中的客户网络随时间不断变化, 而已有对病毒式营销中挖掘技术的研究忽略了网络的动态性, 导致不能获得连续的病毒式营销策略。为了动态挖掘核心群体, 提出一种基于网络更新日志的核心群体动态挖掘算法DMCG, 通过网络更新日志记录更新的节点与事件类型等信息, 在重新选择核心节点时, 根据记录的不同事件类型进行相应处理以避免对未变化节点的重复计算, 从而降低时间消耗。实验表明, DMCG算法能够有效地动态挖掘核心群体, 能较好满足制定分段型病毒式营销策略的需要。
The customer network in E-commerce systems changes continuously over time, while the dynamics is ignored by current researches on data mining for viral marketing, making sequential marketing strategies unavailable.For mine core group, a dynamic mining algorithm called DMCG is proposed, which is based on network change log.By using some necessary information recorded in network change log, such as updated nodes, event type and so on, corresponding processing is performed to avoid re-computing the unchanged nodes when choosing core group in the next step.The experiments results show that DMCG can dynamically mine core group effectively, and meet the needs of making sequential marketing strategies.
[1]中国互联网络信息统计中心.第30次中国互联网络发展状况统计报告[R].北京:2012.
[2]2011中国市场研究报告网[EB/OL].http://www.ewise.com.cn.
[3]夏秀峰, 赵秀涛.客户信任网络下病毒式营销核心群体的挖掘[J].小型微型计算机系统, 2011(1):56-60.
[4]赵秀涛.Web病毒式营销中的挖掘技术研究[D].沈阳:沈阳航空工业学院, 2011.
[5]Mislove A, Marcon M, Gummad K P.2007 Internet Measurement Conference, Proc of the 7th ACM SIGCOMM conference on Internet measurement, [c]San Diego, California, USA, 2007.
[6]JΦsang A, Bhuiyan T.Optimal trust networks analysis with subjective logic[ C]//The second International Conference on Emerging Security Information, Systems and Technologies.France, 2008.
[7]Xia Xiufeng, Li Ying, Cong Lihui.Web viral marketing based construction strategy of customer trust Network[C]//The Fourth International Conference on Computational and Information Sciences, 2012:425-429.
[8]Yang J M, Yao C Z, Ma W C.A study of the spreading scheme for viral marketing based on acomplex network model[J].physica A, 2010 (389):859-870.