基于布朗桥模型的重要同现模式挖掘  

Important Co-occurrence Pattern Mining Based on Brown Bridge Model

作　　者：邓超[1,2] 罗泽[1] 阎保平[1] 

机构地区：[1]中国科学院计算机网络信息中心,北京100190 [2]中国科学院大学,北京100049

出　　处：《计算机工程》2014年第12期63-67,共5页Computer Engineering

基　　金：国家自然科学基金资助项目(中美软件合作研究项目)(61361126011)

摘　　要：为研究动物迁徙过程中的群体行为特点,需要发现动物的群体性停留区域和时间,然而现有同现模式挖掘算法只关注动物群体同现的瞬时性而未关注同现的持续性。为此,结合同现模式挖掘和经停地分析,提出基于布朗桥模型的重要同现模式挖掘算法。利用布朗桥模型对时空对象的轨迹进行建模,得到轨迹对应的经停地,并在相交经停地中,通过Apriori算法得到重要同现模式。应用青海湖斑头雁的时空数据实验证明了该算法的正确性,并通过分析挖掘出的时空同现模式,发现了斑头雁迁徙过程中的群体性起点区域、终点区域和中途经停区域。In order to study characteristic of group behavior during the migration of animals,group stopover regions and time of moving objects need to be found,but existing co-occurrence mining algorithms only focus on animal group co-occurrence instantaneity and do not concern with sustainability. To solve the problem,this paper proposes a algorithm for mining important co-occurrence patterns based on Brown bridge model. It finds stopover regions of objects using Brown bridge model and finds important co-occurrences using an Apriori algorithm in the intersections of stopover regions. After all,an experiment using the spatio-temporal data of bar-headed goose in the Qinghai Lake Area is made to prove correctness of the algorithm. By mining the important co-occurrence patterns in the experiment, the group starting regions,ending regions and stopover regions during the migration of the bar-headed gooses are found by analyzing the spatio-temporal mode.

关 键 词：时空数据挖掘 重要同现模式 布朗桥 经停地区域 轨迹建模 概率模型 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]