计算机集成制造系统 ›› 2016, Vol. 22 ›› Issue (第2期): 302-311.DOI: 10.13196/j.cims.2016.02.003

• 产品创新开发技术 • 上一篇    下一篇

基于融合上下文的移动用户行为过程挖掘与预测

王佳秋,于浩,王忠杰   

  1. 哈尔滨工业大学计算机科学与技术学院
  • 出版日期:2016-02-29 发布日期:2016-02-29

Behavioral process mining and predicting of mobile users based on context-fusion

  • Online:2016-02-29 Published:2016-02-29

摘要: 针对移动环境中单个用户个性化行为过程的挖掘和预测问题,考虑不同类型上下文对行为过程的影响,研究将行为过程中不同类型的上下文融合成统一的整体(情景),提出上下文融合过程模型。进而将上下文融合过程模型融入移动用户的行为过程中,提出一种基于融合上下文的行为过程模型。在此基础上提出一种挖掘算法ASCF-Mine,利用行为过程中存在的情景周期和时间属性来挖掘频繁的行为过程。结合频繁的行为过程,提出一种基于协同过滤的预测方法来自主构建满足用户个性化需求的行为过程。通过具体的实验分析,验证了所提方法的有效性。

关键词: 上下文, 上下文融合过程模型, 行为过程挖掘, 行为过程预测

Abstract: Aiming at the problem that mining and predict personalized behavioral process for user in mobile environment,a Context-Fusion Process (CFP) model  was proposed by integrating various contexts into a unified situation.Furthermore,a behavioral process model based on context-fusion was proposed through CFP merged with behavioral process.On this basis,a method called App Sequence based on Context-Fusion Mine (ASCF-Mine) for mining behavioral process was proposed,which took advantage of situation cycle and time properties of process to mine frequent behavioral processes.A predicting approach was proposed based on collaborative filtering to predict the behavioral process for user,which was combined with frequent behavioral process.Through the specific examination,the effectiveness of proposed methods was validated.

Key words: context, context-fusion process model, behavioral process mining, behavioral process prediction

中图分类号: