人群聚集热点区域分析与预测

发布时间：2018-03-14 06:23

本文选题：移动轨迹数据　切入点：聚类分析　出处：《宁夏大学》2017年硕士论文　论文类型：学位论文

【摘要】：社会的快速发展使得人们的生活越来越便捷,人们对出行的需求也越来越高。出租车因为其便捷,成为人们出行时通常会选择的交通工具之一。在人群聚集的区域,人们通常会遇到打车难的问题,而大部分出租车却因为搭载不到乘客而长时间处于空载状态,于是在人群聚集的区域因为人群疏导不及时经常会出现踩踏事件,同时因为大量的出租车在道路上游荡寻找乘客而造成交通拥堵的问题。GPS轨迹数据的出现能够为解决这些问题提供新的方法和思路,由于移动定位技术的快速发展,使得GPS轨迹数据的途径越来越多。出租车是用来收集GPS轨迹数据最常使用的工具,分析和挖掘出租车移动GPS轨迹数据不仅能够帮助本文了解人群聚集热点区域的分布,也能更加有效地提取和预测出载客热点区域。聚类方法是本文提取日常出行热点区域的主要研究方法。通过仿真实验对比三种在基于轨迹数据的热点区域提取中应用最为广泛的聚类算法:K-Means、DBSCAN和Single-Link,本文选择DBSCAN聚类算法对人群聚集热点区域进行提取和分析。通过对不同日期相同时间段以及相同日期不同时间段人群聚集热点区域的分布和规律进行挖掘后,发现相同时间段之间热点区域存在较高的覆盖率,而不同时间段之间热点区域的区域存在较大的差异,从而通过提取相同时间段热点区域重复的载客点位置数据,并将聚类结果作为预测结果,并计算不同热点区域的推荐度,推荐度越高,则表明载客成功概率越高。最后实现出租车载客热点推荐系统为用户推荐推荐度最高的热点区域,从而提高出租车司机的载客成功概率。
[Abstract]:The rapid development of society makes people's life more and more convenient, and people's demand for travel is also increasing. Because of its convenience, taxi becomes one of the means of transportation that people usually choose when they travel. People usually have trouble getting a taxi, but most taxis are out of traffic for a long time because they can't pick up passengers, so stampedes often occur in crowded areas because the crowds are not moving in time. At the same time, because a large number of taxis loitering on the road looking for passengers causing traffic congestion. The appearance of GPS track data can provide new methods and ideas for solving these problems. Because of the rapid development of mobile positioning technology, Taxi is the most commonly used tool to collect GPS trajectory data. The analysis and mining of taxi mobile GPS trajectory data can not only help this paper understand the distribution of the hot areas where people gather. Clustering method is the main research method to extract the daily travel hot spot area. Through the simulation experiment, three kinds of hot spot area extraction based on trajectory data can be compared. The most widely used clustering algorithms are: K-Meansn DBSCAN and Single-Link. in this paper, we select DBSCAN clustering algorithm to extract and analyze the hot spots of crowd clustering. After mining the distribution and rules of the domain, It is found that there is a high coverage rate of hot spots in the same time period, but there are great differences in the hot spot areas in different time periods, so the repeated passenger location data of hot spots in the same time period can be extracted. The higher the recommendation degree is, the higher the success probability of carrying passengers is. Finally, the hot spot recommendation system of taxi is the hot area with the highest recommendation degree. In order to improve the taxi driver's success probability of carrying passengers.
【学位授予单位】：宁夏大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U495;TP311.13

【参考文献】