基于出租车GPS轨迹的智能推荐系统

发布时间：2018-04-09 00:01

本文选题：GPS轨迹　切入点：数据挖掘　出处：《长安大学》2015年硕士论文

【摘要】：目前,城市中的出租车司机由于盲目的四处寻找乘客而造成了出租车资源的严重浪费,增加了燃油的消耗,即加重了环境污染,又降低了出租车司机的收入。而同时又有大量的乘客因为要到处寻找出租车而给自己的出行造成了不便。这种情况的发生主要是因为出租车司机和乘客的行为是盲目的,缺乏统一的指导,从而给出租车司机和乘客均带来了很大的困扰。如今因为监管和安全应用的需求,城市中的出租车都安装了GPS定位装置,而由此产生了大量的出租车GPS轨迹数据信息,通过对出租车GPS轨迹数据的分析,研究其应用成为一个热点。在此背景下,本文通过对西安市出租车历史轨迹数据挖掘和分析,建立了以提高打车响应概率和最优路径推荐为目的的打车推荐系统,设计并开发了网页客户端交互软件,并进行了软件功能及性能方面的测试。论文具体完成的工作如下:(1)本文设计了一种基于网格的乘客行为聚类算法,通过对分时段固定区域内乘客上下车情况分析,提取出了最容易搭载到乘客/乘客乘车的地点数据,并以此作为依据向出租车司机/乘客推荐行驶路径。(2)根据出租车GPS轨迹数据,分析和提取出租车司机经验路径,建立了以出租车历史GPS轨迹数据和当前轨迹数据结合的打车地点以及当前最优路径推荐算法。(3)使用Oracle数据库和MATLAB软件对出租车GPS轨迹数据进行运算和处理;以新浪云为后台,百度地图API为接口,设计开发了客户端软件,方便用户登录网页查看推荐的热点和路径。
[Abstract]:At present, taxi drivers in cities have caused a serious waste of taxi resources and fuel consumption because of blindly searching for passengers, which has aggravated the environmental pollution and reduced the income of taxi drivers.At the same time, there are a large number of passengers to find taxis everywhere, causing inconvenience to their travel.This happens mainly because the behavior of taxi drivers and passengers is blind and lack of unified guidance, which brings great troubles to both taxi drivers and passengers.Nowadays, because of the need of supervision and safety application, taxis in cities are equipped with GPS positioning devices, and a large amount of taxi GPS locus information is generated, through the analysis of taxi GPS locus data.Research on its application has become a hot spot.Under this background, by mining and analyzing the historical track data of taxi in Xi'an, this paper establishes a taxi recommendation system which aims to improve the response probability and the recommendation of the optimal path, and designs and develops the interactive software of the web client.The software function and performance are tested.The main work of this paper is as follows: (1) this paper designs a grid-based clustering algorithm for passenger behavior. By analyzing the situation of passengers getting on and off in a fixed area of time, we extract the most easily picked up data of passenger / passenger location.According to the GPS track data of the taxi, the empirical path of the taxi driver is analyzed and extracted.In this paper, we set up a taxicab location which combines the taxi history GPS track data with the current track data and the current optimal path recommendation algorithm. We use the Oracle database and MATLAB software to calculate and process the taxi GPS track data, and take the Sina cloud as the background.Baidu map API interface, the design and development of client software, easy for users to log in to the web page to view the recommended hot spots and paths.
【学位授予单位】：长安大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U495;TP391.3

【参考文献】