一种面向时空数据的关联规则更新算法

An Updating Algorithm for Spatial and Temporal Data Association Rule

现有的关联规则更新算法大多具有产生大量候选项集和多次扫描数据库的弊端,而且对时空数据的研究少之又少。针对此问题,论文提出一种基于滑动窗口的关联规则更新算法,此算法将访问数据进行行程长度编码并存储于存储器中,然后只需对存储器中的编码数据进行挖掘,不需反复读取数据库信息。同时该算法在由频繁项集产生候选项集时添加了空间约束条件,过滤了空间不相关数据,提高了算法的执行速度和处理效能。通过实验论证,此算法具有更高的挖掘效率,对智能交通、指挥控制等领域有着重要的应用价值。

Most of the present updating association rule algorithms have drawbacks that produce a large number of candidate sets,multiple scans of the database,and have a little research on the spatial and temporal data.To solve this problem,an updating association rule algorithm based on sliding window is proposed in this paper which encodes access data in memory and then only mines the encoding data in memory directly,without repeatedly reading the database information.Meanwhile,the algorithm adds a space constraints to filter irrelevant space data when generating candidate sets by frequent itemsets to improve the execution speed and processing performance.Experiment results show that the algorithm has higher mining efficiency and has important application value for intelligent transportation,command and control,etc.