OSAF-tree--可迭代的移动序列模式挖掘及增量更新方法

移动通信技术和无限定位技术的发展积累了海量的、动态增长的时空数据．利用数据挖掘技术从移动用户的时空行为轨迹当中挖掘用户移动序列模式,在移动通信、交通管理、基于位置服务等领域有着广泛的应用前景．由于移动环境网络资源珍贵、数据量大的特点,传统的序列模式挖掘方法在效率上很难满足需求．OSAF-tree算法基于投影的概念,只需要对数据库进行一遍扫描,就可以很好地处理移动序列模式的挖掘及其增量更新和迭代挖掘问题,这是一个非常高效的算法．与已有的方法相比,OSAF-tree算法在性能和I/O代价等方面都具有明显的优势．

OSAF-tree--An Interactive and Incremental Algorithm for Moving Sequential Pattern Mining

Advances in mobile communication and location determination technology have resulted in a mass of spatio temporal data of mobile users Moving sequential patterns mined from saptio temporal data effectively and efficiently can be used to enhance the performance of the mobile communication network and support decision making for location based services, intelligent transportation systems, etc However, with the rare network resources and massive mobile data, the traditional sequential pattern mining methods are not efficient enough An efficient project based algorithm, OSAF tree, is presented for moving sequential pattern mining This algorithm can get the maximal sequential patterns with one scan of databases and thereby is much more efficient What's more, with a materialized OSAF tree structure, OSAF tree algorithm avoids scanning the database from scratch so that it supports interactive and incremental mining of moving sequential pattern with lower cost Experiments show that OSAF tree algorithm gains a great advantage over algorithms before