Short-term Passenger Flow Prediction of Urban Rail Transit Based on LSTM-KAN-Stacking Combination Model
DOI:
https://doi.org/10.54097/3jn7ak36Keywords:
Urban rail transit, Short-term passenger flow prediction, KAN model, Feature stacking, Combined prediction models.Abstract
Urban rail transit passenger flow prediction is crucial for intelligent management and service of rail transit. Existing research has made some progress in improving passenger flow prediction accuracy, but there is still room for improvement in prediction accuracy. In addition, as the time granularity decreases, the complexity and uncertainty of passenger flow data increase, and the prediction accuracy also decreases. In order to further improve the short-term passenger flow prediction accuracy of rail transit, introduced the KAN network into the field of rail transit passenger flow prediction, and proposed a combined LSTM-KAN-Stacking model. The innovations of this combined model are mainly reflected in two aspects: first, KAN network is integrated into LSTM as a fully connected layer, and the input sequence data enters into KAN after extracting features by LSTM, which enhances the ability to capture complex patterns and improves the model prediction accuracy and robustness. Secondly, the deep learning model is combined with the Stacking integrated learning model, and the LSTM-KAN predicted output values are added to the original time series data as new features to provide richer information input for the Stacking model. By constructing the LSTM-KAN-Stacking model, feature stacking after predicting the original time series with LSTM-KAN model, constructing the integrated model Stacking and outputting the prediction after parameter optimization and relevance assessment, and finally conducting experimental comparative analysis based on Hangzhou metro passenger flow data, the validity of the proposed model is verified. The results show that the LSTM-KAN-Stacking model can reduce the MAE by at least 0.5%, 2.4%, and 5.6% compared with the nine base models, LSTM, LSTM-KAN, and Stacking model, respectively, on the 15-min, 30-min, and 60-min granularity datasets. In addition, the proposed model's goodness-of-fit at three temporal granularities can reach 0.972, 0.973, and 0.941, respectively. Visualization analysis verifies its generalization ability in different types of sites and different temporal patterns. Therefore, LSTM-KAN-Stacking has higher prediction accuracy compared with the traditional model, and KAN network has potential for urban rail transit short-term passenger flow prediction, especially at finer granularity, with high stability and applicability.
Downloads
References
[1] CHUWANG D D, CHEN Weiya, ZHONG Ming. Short-term Urban Rail Transit Passenger Flow Forecasting Based on Fusion Model Methods Using Univariate Time Series[J]. Applied Soft Computing, 2023, 147: 110740.
[2] ZHU Haining. N days average volume based ARIMA forecasting model for Shanghai metro passenger flow[C]//2010 International Conference on Artificial Intelligence and Education (ICAIE). IEEE, 2010: 631-635.
[3] ZHANG Guoyun, JIN Hui. Research on Short-term Passenger Flow Prediction of Urban Rail Transit Based on Improved ARIMA Model[J]. Computer Applications and Software,2022,39(01):339-344.
[4] OKUTANI I, STEPHANEDES Y J. Dynamic prediction of traffic volume through Kalman filtering theory[J]. Transportation Research Part B: Methodological, 1984, 18 (1): 1−11.
[5] DA SILVA D, SHALABY A. Forecasting short-term subway passenger flow using Wi-Fi data: comparative analysis of advanced time-series methods[J]. Journal of Intelligent Transportation Systems, 2024: 1-21.
[6] LI Shuqing, LI Wei, LIU Yaohong, et al. Short-term Passenger Flow Prediction Model of Rail Transit Based on Combined Deep Learning[J]. Journal of Chongqing Jiaotong University (Natural Science Edition),2024,43(02):92-99.
[7] LI Chuan, WANG Xiaodan, CHENG Zhiwei, et al. Forecasting bus passenger flows by using a clustering-based support vector regression approach[J]. Ieee Access, 2020, 8: 19717-19725.
[8] XU Xinyue, KONG Qingxue, LI Jianmin, et al. Analysis of the Spatiotemporal Heterogeneity of the Built Environment on the Passenger Flow of Rail Transit[J]. Journal of Transportation Systems Engineering and Information,2023,23(04):194-202+281.
[9] SHI Zhi, ZHANG Na, SCHONFELD P M, et al. Short-term metro passenger flow forecasting using ensemble-chaos support vector regression[J]. Transportmetrica A: Transport Science, 2020, 16(2): 194-212.
[10] CHEN Quanchao, WEN Di, LI Xuqiang, et al. Empirical mode decomposition based long short-term memory neural network forecasting model for the short-term metro passenger flow[J]. PloS one, 2019, 14(9): e0222365.
[11] ZHANG Guoyun, JIN Hui. Passenger Flow Prediction of Urban Rail Transit Station Based on Improved PSO-LSTM Model[J]. Journal of Computer Applications and Software,2021,38(12):110-114+134.
[12] ZHANG Huizhen, GAO Zhengkai, LI Jianqiang, et al. Short-term Passenger Flow Prediction of Urban Rail Transit Based on Recurrent Neural Network[J]. Journal of Jilin University (Engineering Science),2023,53(02):430-438.
[13] LI Xin, ZHANG Xu, YU Le'an, et al. Research on Short-term Passenger Flow Prediction of Scenic Area Based on Improved Transformer Model[J]. Chinese Management Science, 2025,33(02):105-117.
[14] ZHANG Xinyu, SHI Tianyun, LI Haoguang, et al. Collaborative prediction of passenger flow at multiple entry gates of a large passenger station based on CLTAttention[J]. Journal of Railway Science and Engineering, 1-13.
[15] TANG Zhengyi, HUANG Jiahuan, WANG Jinshui, et al. Passenger flow prediction in subway based on adaptive diffusion graph convolutional attention network[J]. Journal of Railway Science and Engineering, 2024, 21 (12): 4910-4923.
[16] TANG Jiqiang, YANG Luqi, YANG Wu. Prediction of Short-term Passenger Flow Starts and Ends of Rail Transit Based on Long Short-Term Memory Network and Graph Convolutional Network[J]. Journal of Chongqing University,2022,45(11):91-99.
[17] WANG Jinshui, OU Xuewen, CHEN Junyan, et al. Classification and Prediction of Short-term Passenger Flow of Urban Rail Stations Based on Combined Model[J]. Journal of Railway Science and Engineering,2023,20(06):2004-2012.
[18] YANG Jing, ZHU Jingwei, LIU Bo, et al. Short-term Passenger Flow Prediction of Urban Rail Transit Based on Combined Model[J]. Journal of Transportation Systems Engineering and Information Technology,2019,19(03):119-125.
[19] Zhao Xiaoli. Research on the application of ARIMA model in short-term passenger flow prediction of urban rail transit[J]. Modern Urban Rail Transit,2023, (08):77-82.
[20] Wu Zhenya. Data Analysis and Prediction of Passenger Flow of Shanghai Hongqiao Hub Based on ARIMA Model[J]. Traffic & Transportation,2022,35(S1):315-320)
[21] Williams B M, Hoel L A. Modeling and Forecasting Vehicular Traffic Flow as a Seasonal ARIMA Process: Theoretical Basis and Empirical Results[J]. Journal of Transportation Engineering, 2003, 129(6): 664-672.
[22] Li Jie,Peng Qiyuan,Yang Yuxiang. Passenger Flow Prediction of Guangzhou-Zhuhai Intercity Railway Based on SARIMA Model[J]. Journal of Southwest Jiaotong University,2020,55(01):41-51.
[23] Duan Ran,Pang Jianhua,Zhang Liangjun. Research on Passenger Flow Prediction of Railway Station Based on SARIMA Model[J]. Practice and Understanding of Mathematics,2019,49(09):1-10.
[24] Milenković M, Švadlenka L, Melichar V, et al.SARIMA Modelling Approach for Railway Passenger Flow Forecasting[J]. Transport,2015,1-8.
[25] Han Hao, Xu Sheng'an, Zhao Meng. Short - Term Passenger Flow Prediction Model of Rail Transit Based on LightGBM Considering Network Structure[J]. Railway Transport and Economy, 2021, 43(10): 109 - 117.
[26] Cheng Y, Li H, Sun S, et al.Short-term Subway Passenger Flow Forecasting Approach Based on Multi-source Data Fusion[J]. Information Sciences,2024,679121109-121109.
[27] Yao K, Gao G, Liu Y, et al. A Stable Passenger Flow Forecast Approach for Newly Opened Metro Stations Based on Multi-Source Data and Random Forest Regression Model[C]//2022 3rd International Conference on Intelligent Design (ICID). IEEE, 2022: 249-254.
[28] Yang Xinmiao, Wang Wei, Gu Weiping, et al. Fuzzy Neural Network Prediction Model for Passenger Flow of Bus Lines[J]. Journal of Highway and Transportation Research and Development, 2000, (04):38 - 40 + 46.
[29] Jiang Ping, Shi Qin, Chen Wuwei, et al. Bus Passenger Flow Prediction Based on Elman - type Recurrent Neural Network[J]. Journal of Hefei University of Technology (Natural Science), 2008, (03):340 - 342 + 369.
[30] Zou Dong, Liu Qiong, Huang Zirong. Research on Passenger Flow Prediction of Urban Rail Transit Network Considering Spatio-temporal Correlation[J]. Urban Mass Transit, 2016, 19(03):32 - 37.
[31] Li Chunxiao, Li Haiying, Jiang Xi, et al. Short-Term Station Inbound Passenger Flow Prediction Based on Generalized Dynamic Fuzzy Neural Network[J]. Urban Rapid Rail Transit, 2015, 28(04):57 - 61.
[32] Wang Jinfeng, Sun Lianying, Zhang Tian, et al. Short-Term Passenger Flow Prediction of Urban Rail Transit Based on K-LSTM-ecm Model[J]. Manufacturing Automation, 2022, 44(05):103 - 107 + 133.
[33] Yang Jing, Dai Shengxu, Zhang Hongliang, et al. Short-Term Inbound Passenger Flow Prediction of Subway Station During the End of Large-Scale Events[J]. Science Technology and Engineering, 2021, 21(05):2042 - 2048.
[34] Zhao Liqiang, Li Ruisen, Tang Shuixiong, et al. Short-Term Passenger Flow Prediction of Urban Rail Transit Based on TCN-LSTM with Multidimensional Predictable Features[J]. Journal of Beijing University of Chemical Technology (Natural Science Edition), 2024, 51(05):86 - 96.
[35] Tang Zhengyi, Huang Jiahuan, Wang Jinshui, et al. Subway Passenger Flow Prediction Based on Adaptive Diffusion Graph Convolutional Attention Network[J]. Journal of Railway Science and Engineering, 1 - 13.
[36] Wang Xueqin, Xu Xinyue, Wu Yuankai, et al. Short-Term Passenger Flow Prediction of Urban Rail Transit Based on Hybrid Deep Learning Model[J]. Journal of Railway Science and Engineering, 2022, 19(12):3557 - 3568.
[37] Ma Qian, Liang Yi, Duan Yi, et al. Metro Passenger Flow Prediction Model Based on Dynamic Sparse Attention[J]. Urban Mass Transit, 2022, 25(04):22 - 26.
[38] Zhao Mingwei, Zhang Wensheng. Short - term Passenger Flow Prediction of Urban Rail Transit Based on IPSO - LSTM Combined Model[J]. Railway Transport and Economy, 2022, 44(02):123 - 130.
[39] Liu Jie. Research on Short-Term Prediction Model of Railway Passenger Flow Based on Ensemble Algorithm[J]. Journal of Chongqing Jiaotong University (Natural Science), 2022, 41(05):20 - 25.
[40] Shi Z, Zhang N, Schonfeld P M, et al. Short-term Metro Passenger Flow Forecasting Using Ensemble-chaos Support Vector Regression[J]. Transportmetrica A: Transport Science, 2020, 16(2): 194-212.
[41] Wang X, Zhu C, Jiang J. A Deep Learning and Ensemble Learning Based Architecture for Metro Passenger Flow Forecast[J]. IET Intelligent Transport Systems, 2023, 17(3): 487-502.
[42] DING Xinyu, SHI Yan. Prediction of Short-term Traffic Flow in Road Network Based on Deep Spatiotemporal Residual Network[J]. Computer Engineering and Design,2022,43(02):316-322.
[43] ZENG Lu, JIANG Zixuan, PENG Dongliang, et al. Passenger flow prediction of urban rail transit stations based on adaptive spatio-temporal hierarchical network[J]. Journal of Railway Science and Engineering, 1-13.
[44] LIU Ziming, WANG Yixuan, VAIDYA S, et al. Kan: Kolmogorov-arnold networks[J]. arXiv preprint arXiv:2404.19756, 2024.
[45] WOLPERT D H. Stacked Generalization[J]. Neural networks, 1992, 5(2): 241-259.
[46] YANG Anan, HAN Xingyu, TIAN Kuang, et al. Research on Short-Term Passenger Flow Prediction of Airport Line Based on Stacking Ensemble Learning[J]. Shandong Science, 2024, 37(04):112 - 120.
[47] CHEN Peng, FU Xin, WANG Xiang. A Graph Convolutional Stacked Bidirectional Unidirectional-LSTM Neural Network for Metro Ridership Prediction[J]. IEEE Transactions on Intelligent Transportation Systems, 2021, 23(7): 6950-6962
[48] WANG Zitong, LI Chenliang. Traffic Flow Prediction Based on Location-Aware Spatio-Temporal Graph Convolutional Network[J]. Data Analysis and Knowledge Discovery, 2025,9(02):94-105.
[49] LI Beibei, CHEN Chunxiao, ZHENG Liaomo, et al. Tool Wear Prediction Based on FIR-Stacking[J]. Modular Machine Tool & Automatic Manufacturing Technique, 2024(4):87 - 91.
[50] ZHAO Youlin, PANG Hangyuan, SHI Yanqing. Construction of Health Information Portrait and Identification of False Health Information: Integrating Social Sensing Data and Publisher's Prior Knowledge[J]. Documentation, Information & Knowledge, 2024,41(06):141-154+165.
[51] XIE Qianwei, XUE Fengchang, CHEN Jianfei. Integrated Learning Prediction Model for Rape Flowering Period Combined with Virtual Sample Generation[J]. Transactions of the Chinese Society of Agricultural Engineering, 2024,40(19):159-167.
[52] CHEN Fei, CHEN Zhenjie, LI Feixue, et al. Spatio-temporal Prediction Stacking-GWR Model of Urban Land Price[J]. Geography and Geo-Information Science, 2024, 40(05):1-10.
[53] GUO Zhiqiang, ZHANG Botao, ZENG Yunliu. Research on Kiwifruit Sugar Content Detection by Near-Infrared Spectroscopy Combined with Stacking Ensemble Learning[J]. Spectroscopy and Spectral Analysis, 2024, 44(10):2932-2940.
[54] WANG Xiaoming, XU Bin, YIN Yuanya, et al. Wind Power Prediction Method Based on Spatial Correlation and Stacking Ensemble Learning[J]. Electric Power Engineering Technology, 2024, 43(05):224-232.
Downloads
Published
Issue
Section
License
Copyright (c) 2025 Highlights in Science, Engineering and Technology
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
