摘要
随着大数据时代的到来,人工智能、物联网等技术的发展,移动对象位置信息的急剧膨胀,信息数据已经复杂化和多样化起来,使得研究时空轨迹数据变得相当困难.因此,挖掘复杂多样的时空轨迹数据隐含信息已经成为数据挖掘领域的核心问题.研究发现时空轨迹数据具有分布不均匀和频率不稳定等特点,通过探究其构成方式,提出一种适合时空轨迹数据的预处理方式;同时,基于移动对象在运动中轨迹数据的时空变化,探寻研究对象的活动规律,并以此构建了热点区域发现和周期模式发现方法;最后,从实际应用的角度出发,采用基于划分和基于密度的聚类算法,以2016至2019年白头鹎观测数据为研究对象,对其运动的热点区域进行检测,实验结果与白头鸨真实活动轨迹相符,证明方法有效,可实现对移动对象活动规律的探索.
With the rapid development and application of Internet technology and mobile intelligence technology,the rapid expansion of the location information of mobile objects also makes the information content develop towards the direction of diversification and complexity,which also brings great difficulties to the research of spatio-temporal trajectory data.Therefore,the core problem and key in the field of spatio-temporal data mining has gradually changed to how to efficiently mine the implied valuable information from these complex and multifaceted spatio-temporal trajectory data.This paper takes the spatio-temporal characteristics of the spatio-temporal trajectory data with unstable sampling frequency and sparse trajectory points as the starting point,delves into the spatio-temporal composition of the trajectory and designs a pre-processing method suitable for the spatio-temporal trajectory data.At the same time,based on the spatial and temporal changes of the trajectory data of moving objects in motion,the activity pattern of the research object is explored,and the hot spot area discovery and periodic pattern discovery methods are constructed in this way.Finally,from the perspective of practical application,the Chinese Bulbul observation data from 2016 to 2019 were taken as the research object to realize the algorithm proposed in this paper,combining the partitioning based clustering algorithm and the density based clustering algorithm,to detect the hot spots of moving targets.The experimental results are consistent with the real activity track of the Chinese Bulbul,which proves that the method is effective,so as to realize the exploration of the activity law of the mobile object.
作者
陆妍玲
韦晶闪
赵雨萌
周俊芬
李景文
姜建武
LU Yan-ling;WEI Jing-shan;ZHAO Yu-meng;ZHOU Jun-fen;LI Jing-wen;JIANG Jian-wu(Guilin University of Technology,Guilin 541004,China;Guangxi Key Laboratory of Spatial Information and Geomatics,Guilin 541004,China)
出处
《数学的实践与认识》
2021年第13期129-138,共10页
Mathematics in Practice and Theory
基金
国家自然科学基金(地区基金)(41961063)
广西空间信息与测绘重点实验室主任基金(桂科能16-380-25-17、桂科能 15-140-07-14)。
关键词
数据挖掘
轨迹聚类
热点区域
活动周期
moving object
spatio-temporal trajectory
hot-region
cycle mode
