Spark平台上基于K-Means算法的热点路径发现方法研究

发布时间：2018-01-17 16:15

本文关键词：Spark平台上基于K-Means算法的热点路径发现方法研究　出处：《信息工程大学学报》2016年06期 　论文类型：期刊论文

【摘要】：移动对象的热点路径发现问题是大数据环境下位置服务研究的一个热点,可支撑用户行程推荐、智能交通管控以及城市道路规划等诸多应用。当前,随着智能移动终端的快速发展和应用,移动对象轨迹数据的规模量日益增大,而现有的轨迹热点路径发现方法在大数据环境下存在处理效率不高的问题。针对该问题,提出了在Spark平台上基于K-means聚类算法的热点路径发现方法。该方法首先确定轨迹中的停留点,通过停留点将轨迹划分为若干轨迹段,给出轨迹段之间距离的定义和计算方法,在此基础上采用K-means算法,对移动轨迹段进行聚类,聚类结果即反映出轨迹中的热点路径。实验结果表明,在Spark平台上实现该方法有效的提高了在大规模数据集上发现热点路径的效率。
[Abstract]:Hot path of moving objects found that the problem is a hot research topic in location services under the big data environment, and can support the user travel recommendation, intelligent traffic control and city road planning and many other applications. At present, with the rapid development and application of intelligent mobile terminal, the mobile object scale trajectory data is increasing, and the trajectory of hot path the existing discovery method has the problem of low efficiency in large data environment. To solve this problem, proposed on the platform of Spark hot path clustering algorithm based on K-means method. This method firstly determined to stay in the path, stay will be divided into several trajectory trajectory trajectory is given the distance between. The definition and calculation method on the basis of K-means algorithm to cluster the mobile trajectory, the clustering results reflected the hot path. The experimental results It shows that the implementation of this method on the Spark platform effectively improves the efficiency of finding hot spots on a large dataset.

【作者单位】：数学工程与先进计算国家重点实验室;
【分类号】：TP311.13
【正文快照】： 近年来,随着卫星定位导航系统的日益成熟,以及Web GIS、移动互联网、移动智能终端等技术的快速发展和广泛应用,位置服务已经成为当前产业界和学术界研究的热点,热点路径的发现是位置服务中的关键内容之一。热点路径是指一段时间内移动对象频繁经过的路线,可以很大程度上反映移

【参考文献】