基于Spark的分布式交通流数据预测系统

在大数据时代,在城市复杂交通环境中,实现实时、准确的交通流预测,是实现智能交通系统的必要前提。提出了一种在Spark平台上基于梯度优化决策树的分布式城市交通流预测模型（distributed urban traffic prediction with GBDT,DUTP-GBDT）;并提出了分布式情况下梯度优化决策树模型实现的优化方法,包括切分点抽样、特征装箱和逐层训练三种,提高了分布式情况下梯度优化决策树训练效率。基于Spark分布式计算平台高效、可靠、弹性可扩展的优势,以及梯度优化决策树模型准确率较高和时间复杂度较低的优点,利用时间特征、道路状况特征以及天气特征等特征参数,建立了DUTP-...

Full description

Saved in:
Bibliographic Details
Published in计算机应用研究 Vol. 35; no. 2; pp. 405 - 409
Main Author 黄廷辉;王玉良;汪振;崔更申
Format Journal Article
LanguageChinese
Published 桂林电子科技大学计算机与信息安全学院,广西桂林541004 2018
桂林电子科技大学广西可信软件重点实验室,广西桂林541004%桂林电子科技大学计算机与信息安全学院,广西桂林,541004
Subjects
Spark平台
交通流预测
分布式计算
梯度优化决策树模型
traffic flow forecast
梯度优化决策树模型
分布式计算
Spark平台
Spark platform
gradient optimization decision tree model
交通流预测
distributed computing
Online AccessGet full text
ISSN1001-3695
DOI10.3969/j.issn.1001-3695.2018.02.019

Cover

More Information
Summary:在大数据时代,在城市复杂交通环境中,实现实时、准确的交通流预测,是实现智能交通系统的必要前提。提出了一种在Spark平台上基于梯度优化决策树的分布式城市交通流预测模型（distributed urban traffic prediction with GBDT,DUTP-GBDT）;并提出了分布式情况下梯度优化决策树模型实现的优化方法,包括切分点抽样、特征装箱和逐层训练三种,提高了分布式情况下梯度优化决策树训练效率。基于Spark分布式计算平台高效、可靠、弹性可扩展的优势,以及梯度优化决策树模型准确率较高和时间复杂度较低的优点,利用时间特征、道路状况特征以及天气特征等特征参数,建立了DUTP-GBDT模型,实现了实时、准确的交通流预测。通过与GABP、GA-KNN、MSTAR等模型的对比,证明了利用Spark平台,DUTP-GBDT模型在分布式环境下准确率和训练速度方面均有所提高,符合城市交通流预测系统的各项要求。
Bibliography:Huang Tinghuia,b, Wang Yuliang1, Wang Zhena, Cui Gengshen1 （ a. School of Computer Science ＆ Information Security, b. Gnangxi Key Laboratory of Trusted Software, Guilin University of Electronic Technology, Guilin Guangxi 541004, China）
51-1196/TP
traffic flow forecast; distributed computing; Spark platform; gradient optimization decision tree model
In the era of big data and complex urban traffic environment, real-time and accurate traffic flow forecast is a prere- quisite tO implementing intelligent transportation system. This paper presented a distributed urban traffic flow forecasting model which based on gradient optimization decision tree on Spark platform. It also proposed the optimization method of gradient optimization decision tree model in distributed case, which included sampling points, feature packing and layer-by-layer training. All of them could improve the training efficiency of gradient optimization decision tree in distributed case. The characteristics of time ,road condition and weather were e
ISSN:1001-3695
DOI:10.3969/j.issn.1001-3695.2018.02.019