Short-term metro passenger flow prediction based on EMD-LSTM
Article Text (Baidu Translation)
-
摘要: 为降低样本噪声对客流预测模型的干扰, 结合深度学习理论, 提出了一种基于经验模态分解与长短时记忆神经网络的短时地铁客流预测模型; 将预测过程分为3个阶段, 第1阶段预处理原始地铁刷卡数据, 构建进(出)站客流时间序列, 运用经验模态分解法将时间序列转化为一系列本征模函数及残差, 第2阶段利用偏自相关函数确定长短时记忆神经网络的输入变量, 第3阶段基于深度学习库Keras, 完成长短时记忆神经网络的搭建、训练及预测; 以上海地铁2号线人民广场站客流数据验证了模型的有效性。计算结果表明: 与代表性的预测模型(差分自回归移动平均模型、支持向量机、经验模态分解与反向传播神经网络、长短时记忆神经网络)相比, 经验模态分解与长短时记忆神经网络预测模型分别将工作日高峰、平峰、全日的进(出)站客流预测精度分别至少提升了2.1%(2.5%)、2.7%(3.5%)、2.7%(3.4%), 将非工作日全日的进(出)站客流预测精度至少提升了3.3%(3.5%), 说明经验模态分解与长短时记忆神经网络的组合是一种预测短时地铁客流的有效模型; 当预测步长由5 min逐渐增加至30 min时, 工作日高峰、平峰和全日进(出)站客流的平均绝对百分比预测误差分别由14.8%(13.9%)、16.8%(17.4%)和16.6%(17.0%)逐渐降低至7.0%(6.2%)、8.3%(7.5%)和8.1%(7.4%), 说明该方法预测误差与预测步长呈负相关。Abstract: To weaken the interference of sample noise to the prediction model of passenger flow, a short-term metro passenger flow prediction model was proposed based on the deep-learning theory, empirical mode decomposition(EMD) and long short-term memory neural network(LSTMNN). The prediction process was divided into three stages.In the first stage, the raw automatic fare collection(AFC) data were preprocessed, and the tap-in(tap-out) passenger flow time series were constructed and decomposed into a series of intrinsic mode functions(IMFs) and a residues by the EMD. In the second stage, the input variables of the LSTMNN were determined by the partial autocorrelation function(PACF). In the third stage, the LSTMNN was developed, trained and predicted through the deep learning library Keras. A case study of Shanghai People's Square Station on metro line 2 was conducted to validate the model performance. Calculation result shows that, compared to the representative prediction models(differential autoregressive integrated moving average model, support vector machine, empirical mode decomposition and back propagation neural network, and LSTMNN), the EMD-LSTM prediction model increases the weekdays' tap-in(tap-out) passenger flow prediction accuracy of peak hour, off-peak hour, and full-day by at least 2.1%(2.5%), 2.7%(3.5%), and 2.7%(3.4%), respectively, and also increases the weekends' tap-in(tap-out) passenger flow prediction accuracy of full-day by at least 3.3%(3.5%). Thus, the EMD-LSTM is effective to predict the short-term metro passenger flow. When the forecasting step gradually increases from 5 minutes to 30 minutes, the weekdays' tap-in(tap-out)average absolute percentage prediction errors of peak hour, off-peak hour, and full-day gradually decreases from 14.8%(13.9%), 16.8%(17.4%), and 16.6%(17.0%) to 7.0%(6.2%), 8.3%(7.5%), and 8.1%(7.4%), respectively. Therefore, the forecasting error of EMD-LSTM is negatively correlated with the forecasting step length.
-
图 3 日进站客流序列EMD分解结果
Figure 3. EMD decomposition results of daily tap-in passenger flow sequence
图 4 本征模函数分量和残差的偏自相关系数
Figure 4. Partial autocorrelation coefficients of IMFs and residue
表 1 预测误差对比
Table 1. Comparison of prediction errors
客流类型 模型 工作日 非工作日 高峰 平峰 全日 全日 E1/% E2 E1/% E2 E1/% E2 E1/% E2 进站 ARIMA 15.23 695 17.17 237 16.93 308 18.43 320 SVM 14.21 642 16.40 205 16.13 274 17.62 298 LSTM 12.41 540 15.39 189 15.03 264 16.05 271 EMD-BPNN 10.75 473 13.49 175 13.16 248 14.56 254 EMD-LSTM 8.62 368 10.75 161 10.49 211 11.23 198 出站 ARIMA 14.61 728 16.92 251 16.64 302 17.68 310 SVM 13.29 690 16.49 220 16.10 298 17.60 308 LSTM 11.92 659 15.54 204 15.10 291 16.39 289 EMD-BPNN 9.76 562 13.07 164 12.67 246 14.62 261 EMD-LSTM 7.31 421 9.56 117 9.29 175 11.08 178 
下载: 导出CSV
-
[1] 蔡昌俊, 姚恩建, 王梅英, 等. 基于乘积ARIMA模型的城市轨道交通进出站客流量预测[J]. 北京交通大学学报, 2014, 38(2): 135-140. https://www.cnki.com.cn/Article/CJFDTOTAL-BFJT201402024.htmCAI Chang-jun, YAO En-jian, WANG Mei-ying, et al. Prediction of urban railway station's entrance and exit passenger flow based on multiply ARIMA model[J]. Journal of Beijing Jiaotong University, 2014, 38(2): 135-140. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-BFJT201402024.htm [2] 袁坚, 王鹏, 王钺, 等. 基于时空特征的城市轨道交通客流量预测方法[J]. 北京交通大学学报, 2017, 41(6): 42-48. https://www.cnki.com.cn/Article/CJFDTOTAL-BFJT201706009.htmYUAN Jian, WANG Peng, WANG Yue, et al. A passenger volume prediction method based on temporal and spatial characteristics for urban rail transit[J]. Journal of Beijing Jiaotong University, 2017, 41(6): 42-48. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-BFJT201706009.htm [3] 杨军, 侯忠生. 一种基于灰色马尔科夫的大客流实时预测模型[J]. 北京交通大学学报, 2013, 37(2): 119-123, 128. doi: 10.3969/j.issn.1673-0291.2013.02.022YANG Jun, HOU Zhong-sheng. A grey Markov based on large passenger flow real-time prediction model[J]. Journal of Beijing Jiaotong University, 2013, 37(2): 119-123, 128. (in Chinese). doi: 10.3969/j.issn.1673-0291.2013.02.022 [4] 王兴川, 姚恩建, 刘莎莎. 基于AFC数据的大型活动期间城市轨道交通客流预测[J]. 北京交通大学学报, 2018, 42(1): 87-93. https://www.cnki.com.cn/Article/CJFDTOTAL-BFJT201801013.htmWANG Xing-chuan, YAO En-jian, LIU Sha-sha. Urban rail transit passenger flow forecasting for large special event based on AFC data[J]. Journal of Beijing Jiaotong University, 2018, 42(1): 87-93. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-BFJT201801013.htm [5] 姚恩建, 周文华, 张永生. 城市轨道交通新站开通初期实时进出站客流量预测[J]. 中国铁道科学, 2018, 39(2): 119-127. doi: 10.3969/j.issn.1001-4632.2018.02.15YAO En-jian, ZHOU Wen-hua, ZHANG Yong-sheng. Real-time forecast of entrance and exit passenger flow for newly opened of urban rail transit at initial stage[J]. China Railway Science, 2018, 39(2): 119-127. (in Chinese). doi: 10.3969/j.issn.1001-4632.2018.02.15 [6] LI Yang, WANG Xu-dong, SUN Shuo, et al. Forecasting short-term subway passenger flow under special events scenarios using multiscale radial basis function networks[J]. Transportation Research Part C: Emerging Technologies, 2017, 77: 306-328. doi: 10.1016/j.trc.2017.02.005 [7] SUN Yu-xing, LENG Biao, GUAN Wei. A novel wavelet-SVM short-time passenger flow prediction in Beijing subway system[J]. Neurocomputing, 2015, 166: 109-121. doi: 10.1016/j.neucom.2015.03.085 [8] 李得伟, 颜艺星, 曾险峰. 城市轨道交通进站客流量短时组合预测模型[J]. 都市快轨交通, 2017, 30(1): 54-58, 64. doi: 10.3969/j.issn.1672-6073.2017.01.012LI De-wei, YAN Yi-xing, ZENG Xian-feng. Combined short-term prediction model of station entry flow in urban rail transit[J]. Urban Rapid Rail Transit, 2017, 30(1): 54-58, 64. (in Chinese). doi: 10.3969/j.issn.1672-6073.2017.01.012 [9] 熊杰, 关伟, 孙宇星. 基于Kalman滤波的地铁换乘客流预测[J]. 北京交通大学学报, 2013, 37(3): 112-116, 121. doi: 10.3969/j.issn.1673-0291.2013.03.021XIONG Jie, GUAN Wei, SUN Yu-xing. Metro transfer passenger forecasting based on Kalman filter[J]. Journal of Beijing Jiaotong University, 2013, 37(3): 112-116, 121. (in Chinese). doi: 10.3969/j.issn.1673-0291.2013.03.021 [10] 李春晓, 李海鹰, 蒋熙, 等. 基于广义动态模糊神经网络的短时车站进站客流量预测[J]. 都市快轨交通, 2015, 28(4): 57-61. doi: 10.3969/j.issn.1672-6073.2015.04.012LI Chun-xiao, LI Hai-ying, JIANG Xi, et al. Short-term entrance passenger flow forecast at urban rail transit station based on generalized dynamic fuzzy neural networks[J]. Urban Rapid Rail Transit, 2015, 28(4): 57-61. (in Chinese). doi: 10.3969/j.issn.1672-6073.2015.04.012 [11] DING Chuan, WANG Dong-gen, MA Xiao-lei, et al. Predicting short-term subway ridership and prioritizing its influential factors using gradient boosting decision trees[J]. Sustainability, 2016, 8(11): 1-16. [12] LENG Biao, ZENG Jia-bei, XIONG Zhang, et al. Probability tree based passenger flow prediction and its application to the Beijing subway system[J]. Frontiers of Computer Science, 2013, 7(2): 195-203. doi: 10.1007/s11704-013-2057-y [13] ZHAO Zheng, CHEN Wei-hai, WU Xing-ming, et al. LSTM network: a deep learning approach for short-term traffic forecast[J]. IET Intelligent Transport Systems, 2017, 11(2): 68-75. doi: 10.1049/iet-its.2016.0208 [14] MA Xiao-lei, TAO Zhi-min, WANG Yin-hai, et al. Long short-term memory neural network for traffic speed prediction using remote microwave sensor data[J]. Transportation Research Part C: Emerging Technologies, 2015, 54: 187-197. [15] POLSON N G, SOKOLOV V O. Deep learning for short-term traffic flow prediction[J]. Transportation Research Part C: Emerging Technologies, 2017, 79: 1-17. [16] LYU Yi-Sheng, DUAN Yan-Jie, KANG Wen-wen, et al. Traffic flow prediction with big data: a deep learning approach[J]. IEEE Transactions on Intelligent Transportation Systems, 2015, 16(2): 865-873. [17] LIU Li-juan, CHEN Rung-ching. A novel passenger flow prediction model using deep learning methods[J]. Transportation Research Part C: Emerging Technologies, 2017, 84: 74-91. [18] HUANG Wen-hao, SONG Guo-jie, HONG Hai-kun, et al. Deep architecture for traffic flow prediction: deep belief networks with multitask learning[J]. IEEE Transactions on Intelligent Transportation Systems, 2014, 15(5): 2191-2201. [19] HUANG N E, SHEN Z, LONG S R, et al. The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis[J]. Proceedings of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences, 1998, 454: 903-995. [20] WEI Yu, CHEN Mu-chen. Forecasting the short-term metro passenger flow with empirical mode decomposition and neural networks[J]. Transportation Research Part C: Emerging Technologies, 2012, 21(1): 148-162. [21] CHEN Syuan-yi, CHOU Wei-yao. Short-term traffic flow prediction using EMD-based recurrent Hermite neural network approach[C]∥IEEE. 15th International IEEE Conference on Intelligent Transportation Systems. New York: IEEE, 2012: 1821-1826. [22] WANG Hai-zhong, LIU Lu, DONG Shang-jia, et al. A novel work zone short-term vehicle-type specific traffic speed prediction model through the hybrid EMD-ARIMA framework[J]. Transportmetrica B: Transport Dynamics, 2016, 4(3): 159-186. [23] HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997, 9(8): 1735-1780. [24] ZHONG Chen, BATTY M, MANLEY E D, et al. Variability in regularity: mining temporal mobility patterns in London, Singapore and Beijing using smart-card data[J]. PloS One, 2016, 11(2): 1-17. [25] TANG Li-yang, ZHAO Yang, JAVIER C, et al. Forecasting short-term passenger flow: an empirical study on Shenzhen metro[J]. IEEE Transactions on Intelligent Transportation Systems, 2019, 20(10): 3613-3622. [26] AN Ning, ZHAO Wei-gang, WANG Jian-zhou, et al. Using multi-output feedforward neural network with empirical mode decomposition based signal filtering for electricity demand forecasting[J]. Energy, 2013, 49: 279-288. [27] ZHENG Hui-ting, YUAN Jia-bin, CHEN Long. Short-term load forecasting using EMD-LSTM neural networks with a Xgboost algorithm for feature importance evaluation[J]. Energies, 2017, 10(8): 1-20. [28] ZHANG Xi-ke, ZHANG Qiu-wen, ZHANG Gui, et al. A novel hybrid data-driven model for daily land surface temperature forecasting using long short-term memory neural network based on ensemble empirical mode decomposition[J]. International Journal of Environmental Research and Public Health, 2018, 15(5): 1-23. [29] CHERKSSKY V, MA Y Q. Practical selection of SVM parameters and noise estimation for SVM regression[J]. Neural Networks, 2004, 17(1): 113-126. [30] 张晚笛, 陈峰, 王子甲, 等. 基于多时间粒度的地铁出行规律相似性度量[J]. 铁道学报, 2018, 40(4): 9-17. https://www.cnki.com.cn/Article/CJFDTOTAL-TDXB201804002.htmZHANG Wan-di, CHEN Feng, WANG Zi-jia, et al. Similarity measurement of metro travel rules based on multi-time granularities[J]. Journal of the China Railway Society, 2018, 40(4): 9-17. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-TDXB201804002.htm -
点击查看大图
图(7) / 表(1)
计量
- 文章访问数: 1726
- HTML全文浏览量: 443
- PDF下载量: 848
- 被引次数: 0
