摘要
针对当前城市交通日益复杂脆弱,以及精细化控制对预测精度要求的现实需求,分析短时交通流量预测研究现状及已有方法在实际预测中的特点与局限性,剖析传统预测手段所面临的挑战与困境。研究结合现代城市交通数字化、信息化、智慧化发展背景,把握交通数据由小样本环境向大数据环境转变的有利契机,依据从实际交通大数据中提取的典型数据,分析探讨从海量数据中挖掘具有相似变化态势的数据进行短时交通流量预测的可行性,并从历史数据库构建、相似度量机制、预测相关参数选取等方面提出相应的算法思路和关键技术。
This paper analyzes the features and weaknesses of the traditional methods that have been used to forecast short-term traffic flow. Focusing on the ever-complicated urban traffic and the demand for fine-grained control in improving the precision, the authors have dug into the challengers these traditional approaches have to face. Against the backdrop of the digital and intelligent development of modern urban traffic where the database is in transition from small samples to a big data environment, the paper explores the feasibility of forecasting short-term traffic flow based on the extracted data with similar trends from mass data. Based on the typical extracted data from the actural massive traffic data, this paper puts forward an algorithm and its key technologies in shortterm traffic flow forecasting, which is proved to be of higher precision due to its better performance in simplifing the short-term traffic flow forecasting model.
作者
蔡晓禹
谭宇婷
雷财林
刘秀彩
CAI Xiao-yu, TAN Yu-ting, LEI Cai-lin, LIU Xiu-cai(School of Traffic ＆ Transportation, Chongqing Jiaotong University, Chongqing 400074, Chin)
出处
《铁道运输与经济》
北大核心
2018年第8期88-93,共6页
Railway Transport and Economy
基金
重庆市社会事业与民生保障科技创新专项(cstc2015shms-ztzx30002)
关键词
短时交通流量
预测
研究现状
大数据环境
数据挖掘
Short-term Traffic Flow
Forecasting
Current Researches
Big Data Environment
Data Mining
