摘要
目前,主流的轨迹伴随模式挖掘方法大多是对连续短时间内轨迹的一次挖掘,忽略了前后非连续时间上的关联分析,因此对隐含伴随模式的挖掘不准确。本文对轨迹伴随模式进行了分析,并提出一种结合密度聚类和关联分析的伴随模式分析方法。该方法首先挖掘轨迹数据中的局部模式簇,通过非连续时间片局部模式簇的关联分析,优化挖掘结果。实验结果表明本文方法可以有效地挖掘轨迹中的伴随模式。
At present,the mainstream trajectory adjoint pattern mining methods are usually for short time analysis,and most of them mine trajectory data once,rarely taking into account the relevant analysis between before and after discontinuous t im e,so the implicit adjoint pattern mining is not accurate. This paper analyzes the trajectory adjoint pattern,and ppattern mining method based on density clustering and association analysis. F irs t ly,the local pattern clusters in the trajectory data are mined,and the mining results are optimized by the association analysis of the local pattern clusters in discontinuous timeslices. Experimental results showthat the method can effectively and accurately mine the adjoint model of the trajectory.
出处
《计算机与现代化》
2017年第12期82-87,共6页
Computer and Modernization
关键词
目标轨迹数据
伴随模式挖掘
密度聚类
关联分析
群体运动模式
target trajectory data
adjoint pattern mining
density clustering
association analysis
population movement model
