基于数据挖掘的高效取样方法对手机用户的周期运动模式的研究  

Research of The Periodic Movement Patterns of Mobile Phone Users Based on Efficient Sampling Approach of Data Mining

作　　者：胡臻龙[1] 

机构地区：[1]浙江越秀外国语学院,浙江绍兴312000

出　　处：《科技通报》2013年第11期134-139,156,共7页Bulletin of Science and Technology

基　　金：全国教育信息技术研究"十二五"规划2012年度青年课题(126240628)部分研究成果

摘　　要：在移动商务服务方面,通过特定周期内的客户周期性移动的趋势可以有效地对移动系统资源进行分配,并且提供个性化的基于位置的服务(LBS)。本文对最有前途的周期运动模式进行挖掘,整个过程根据图映射来进行,并提出几种取样方法以提高挖掘效率。首先,从时间区间随机选取一条运动路径作为样本;然后,创建一个独特的路径图结构以储存来自样本的运动路径;最后,应用图遍历算法来鉴定哪些运动模式最有前途。实验结果表明该采样方法在定期最有前途的周期运动模式中执行效率高效、可伸缩性优越。Mobile commerce service cyclical trend of moving by the customers within a specific period can be effectively mobile system resources allocation, and providing personalized location-based services （LBS）. Movement patterns of the most promising cycles excavation, the whole process is carried out according to the graph maps, and put forward several sampling methods to improve the efficiency of mining. First of all, from the time interval randomly select a motion path as a sample; then create a unique path graph structure to store the movement path from the sample; Finally, the application graph traversal algorithm which movement patterns to identify the most promising. Experimental results show that the sampling method in the periodic cycle of the most promising motion mode to perform efficient efficiency, scalability superior.

关 键 词：数据挖掘 取样 序列模式 最有前途的周期运动模式 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]