基于移动交易序列模式的用户行为模式增益挖掘研究

Research on user behavior pattern mining gain based on mobile transaction sequential patterns

在移动运算环境中挖掘用户行为模式是目前数据挖掘应用领域新兴的研究热点课题,通过整合商业事务动态路径发现移动商业序列模式,即基于移动用户行为的移动交易序列信息挖掘模式。基于深度优先搜索与宽度优先策略,对单位增益(利润)与类别商品交易量进行关联挖掘,首先是基于增益模式集成挖掘高增益移动行为序列,其次是依赖频次模式挖掘解决不同类型商务模式的顾客行为序列模式,然后是在移动交易环境下验证评价算法的执行效率。提出基于移动商务行为特征数据信息构造树算法,挖掘移动商务行为数据的品类特征信息,在同类算法结果比较中较高水准地表达发掘移动序列行为模式。实验评价实施中分析用户序列商业行为与绩效比较,通过两个典型算法比较,结果表明构造树算法能在综合移动电子商业环境下发现移动商业交易框架下的用户事务行为路径及用户倾向性的交易行为,便于规划及管理现有的移动电子商务平台。

In mobile computing environments mining the user behavior patterns is emerging research hot topic in data mining applications fields,to discover mobile transaction sequence patterns by integrating the dynamic path of business affairs,namely mining mobile purchasing information of transaction sequential based on mobile user behavior.Based on Depth-First-Search and Breadth-First-Strategy association mining the unit gain（profit）and categories of commodity trading volume,the first is mining integrally high gain on mobile behavior sequence based on customer behavior,the second is solving customer behavior sequence model of different types of business model depended on the frequency mining pattern,and then verify the execution efficiency of the algorithms evaluated in mobile transaction environment.Put forward the constructed tree the mining algorithm based on the features data of mobile commerce behavior,work on category of characteristic information in mobile commerce behavior environment,and explore the expression moving sequential patterns of behavior in algorithm comparison result in a higher level.Comparative analysis of user behavior and performance evaluation sequence commercial experimental implementation,the results show that the constructed tree algorithm can fully represent the client＇s business behavior in the comprehensive mobile electronic business environment,the tree algorithm based on mobile user sequence pattern is more suitable for discovering user transaction behavior path under the framework of mobile business transactions and comprehensive mining user transaction behavior tendency to facilitate the planning and management of existing mobile e-commerce platform.