当前位置>主页 > 期刊在线 > 计算机技术 >

计算机技术2020年21期

面向分布式EMDN-GRU 模型的乘客等待时间预测
白宇,郑永玲,蒋顺英,杨楠
(贵州民族大学 数据科学与信息工程学院,贵州 贵阳 550025)

摘  要:面对移动轨迹大数据难以使用传统数据处理平台进行处理,乘客等待时间难以预测,以及GPS 数据无法明确给出车辆行驶方向的问题。文章提出一种基于Spark 的坐标轴车辆方向判别法,并建立了EMDN-GRU 模型对乘客等待时间进行预测,并且与LSTM、GRU、EMD-LSTM 与EMD-GRU 进行比较。案例研究表明:EMDN-GRU 模型明显优于比较模型,其中MAPE 最少提高了8.183%,最大提高了25.729%;在乘客等待时间预测方面具有良好的效果。


关键词:等待时间;EMD 算法;GRU;Spark;车辆方向



中图分类号:O211.61;TP301.6         文献标识码:A           文章编号:2096-4706(2020)21-0059-08


Passenger Waiting Time Prediction for Distributed EMDN-GRU Model

BAI Yu,ZHENG Yongling,JIANG Shunying ,YANG Nan

(School of Data Science and Information Engineering,Guizhou Minzu University,Guiyang 550025,China)

Abstract:Facing with the problems that it is difficult to use traditional data processing platforms to process big data of moving trajectories,it is difficult to predict the waiting time of passengers,and GPS data cannot clearly give the vehicle driving direction.The article proposes a method for judging the vehicle direction of the coordinate axis based on Spark,and establishes the EMDN-GRU model to predict passenger waiting time,and compares it with LSTM,GRU,EMD-LSTM and EMD-GRU. The case study shows that the EMDN-GRU model is significantly better than the comparison model. The MAPE is increased by at least 8.183% and the largest by 25.729%;it has a good effect on passenger waiting time prediction.

Keywords:waiting time;EMD algorithm;GRU;Spark;vehicle direction


参考文献:

[1] YANG C,CHEN J J. A Scalable Data Chunk Similarity BasedCompression Approach for Efficient Big Sensing Data Processing onCloud [J].IEEE Transactions on Knowledge and Data Engineering,2017,29(6):1144-1157.

[2] BASANTA-VAL P,AUDSLEY N C,WELLINGS A J,etal. Architecting Time-Critical Big-Data Systems [J].IEEE Transactionson Big Data,2016,2(4):310-324.

[3] 夏大文. 基于MapReduce 的移动轨迹大数据挖掘方法与应用研究 [D]. 重庆:西南大学,2016.

[4] LU R Q,JIN X L,ZHANG S M,et al. A Study on BigKnowledge and Its Engineering Issues [J].IEEE Transactions onKnowledge and Data Engineering,2019,31(9):1630-1644.

[5] NUNES J,YASMINA B,DELÉCHELLE E,et al. Imageanalysis by bidimensional empirical mode decomposition [J].Image andVision Computing,2003,21(12):1019-1026.

[6] 齐观德,李石坚,潘遥,等. 基于出租车轨迹数据挖掘的乘客候车时间预测 [C]// 第八届和谐人机环境联合学术会议(HHME2012)论文集PCC. 广州:中国计算机学会多媒体技术专业委员会,2012:18-23.

[7] QI G D,PAN G,LI S J,et al. How Long a Passenger Waitsfor a Vacant Taxi--Large-Scale Taxi Trace Mining for Smart Cities [C]//Proceedings of the 2013 IEEE International Conference on GreenComputing and Communications and IEEE Internet of Things and IEEECyber,Physical and Social Computing.Washington:IEEE ComputerSociety,2013:1029-1036.

[8] XU X J,ZHOU J Y,LIU Y,et al. Taxi-RS:Taxi-HuntingRecommendation System Based on Taxi GPS Data [J].IEEE Transactionson Intelligent Transportation Systems,2015,16(4):(

[9] HWANG R H,HSUEH Y L,CHEN Y T. An effective taxirecommender system based on a spatio-temporal factor analysis model [J].Information Sciences,2015,314:28-40.

[10] QIU Z,LI H Y,HONG S D,et al. Finding Vacant TaxisUsing Large Scale GPS Traces [C]//Web-Age Information Management.IEEE,2014:793-804.

[11] 王诏远,李天瑞,程尧,等. 基于经验分布的打车概率和等待时间预测 [J]. 计算机工程与应用,2015,51(24):254-259.

[12] 陆俊天,孙玲,施佺. 基于门控循环单元神经网络的公交到站时间预测 [J]. 南通大学学报(自然科学版),2020,19(2):43-49.

[13] CHEN C,WANG H,YUAN F,et al. Bus travel timeprediction based on deep belief network with back-propagation [J].Neural Computing and Applications,2020,32(14):10435-10449.

[14] MA J M,CHAN J,RISTANOSKI G,et al. Bus travel timeprediction with real-time traffic information [J].Transportation ResearchPart C:Emerging Technologies,2019,105:536-549.

[15] H E P L ,JIANG G Y,LAM S K,et al. Travel-TimePrediction of Bus Journey With Multiple Bus Trips [J].IEEE Transactionson Intelligent Transportation Systems,2019,20(11):4192-4205.

[16] HUANG N E,SHEN Z,LONG S R,et al. The empiricalmode decomposition and the Hilbert spectrum for nonlinear and nonstationarytime series analysis [J].Proceedings of the Royal Society A:Mathematical,Physical and Engineering Sciences,1998,454(1971):903-995.

[17] 戴邵武,陈强强,刘志豪,等. 基于EMD-LSTM 的时间序列预测方法 [J]. 深圳大学学报(理工版),2020,37(3):265-270.

[18] BIAN D J,QIN S Q,WU W .A hybrid AR-DWT-EMDmodel for the short-term prediction of nonlinear and non-stationary shipmotion [C]//Chinese Control and Decision Conference(CCDC),IEEE,2016:4042-4047.

[19] 姚洪刚,沐年国.EMD-LSTM 模型对金融时间序列的预测 [J/OL]. 计算机工程与应用,2020:1-6(2020-12-02).http://kns.cnki.net/kcms/detail/11.2127.TP.20201202.1130.002.html.

[20] 张可,崔乐. 基于PCA-LSTM 模型的多元时间序列分类算法研究 [J]. 统计与决策,2020(15):44-49.

[21] ZHAO W,YANG H F,LI J Q,et al. NetworkTraffic Prediction in Network Security Based on EMD andLSTM [C]//Proceedings of the 9th International Conference onComputer Engineering and Networks(CENet2019). 湖南:南京中爱教育科技有限公司,2020:830-836.

[22] JIANG T J,ZHOU C G,ZHANG H Q. Time SeriesForecasting with an EMD-LSSVM-PSO Ensemble Adaptive LearningParadigm [C]//Proceedings of the 2018 International Conference onComputational Intelligence and Intelligent Systems.New York:Association for Computing Machinery,2018:44-50.


作者简介:

白宇(1994—),女,汉族,贵州仁怀人,硕士研究生,研究方向:统计学、海量数据统计与分析;

郑永玲(1995—),女,汉族,贵州毕节人,硕士研究生,研究方向:统计学、海量数据统计与分析;

蒋顺英(1996—),女,汉族,贵州兴义人,硕士研究生,研究方向:统计学、海量数据统计与分析;

杨楠(1997—),女,汉族,贵州盘县人,硕士研究生,研究方向:统计学、海量数据统计与分析。