現今的社會，車流量預測變得越來越重要。隨著經濟的快速發展，人們開始追求自己的生活品質。各種物流服務、送餐、專人接送服務日趨完善。據統計，台灣現今每1000人就有344輛車，排名在亞洲的第八名，總車輛數更是來到八百多萬輛。
為了防止交通堵塞，一個良好的交通管理系統是非常需要的，並且近年來隨著深度學習的蓬勃發展，讓AI模型強大的功能加入到交通管理系統之中，實現智慧交通系統(Intelligent Transportation Systems , ITSs)，一個能夠具有即時性，並且高準確度的預測未來的車流量變化是智慧交通系統中非常重要的部分。深度學習模型強大的預測能力使這種短期交通流預測問題得到了更好的結果。然而，大多數研究都是基於深度學習模型的更新，很少討論資料集的前處理方法以及資料集的特殊特徵。
因此，本文探討了不同資料預處理方法對交通流資料集的影響以及資料集日間特徵和日內特徵對深度學習模型特徵提取效果的差異。首先，採用多種資料降噪方法，使資料集更穩定，更容易被深度學習模型和統計模型學習。接下來，透過卷積神經網路（Convolution Neural Network, CNN）提取日間和日內車流模式，然後將提取的特徵導入長短期記憶（Long Short-Term Memory, LSTM）單元以學習交通流的時間序列變化。透過對高速公路情況的真實數據案例研究表明，提取資料中的日內與日間特徵，並對其進行不同比例的合併後作為CNN-LSTM輸入特徵，在本研究所設定之預測任務比起不適合使用特徵的模型與使用單一特徵的模型相比，有更低的誤差，更好的預測能力，該方法的錯誤率可以達到5%以下。

Traffic flow prediction in today ‘s society has become more important day by day. With the high-speed development of economy, people start pursuing their own quality of life. Various logistics services, meal delivery and personal pick-up services are improving with each passing day. According to statistics, Taiwan now has 344 vehicles per 1,000 people, ranking eighth in Asia, and the total number of vehicles has reached more than 8 million.
To prevent traffic jams, a good traffic management system is very necessary, and with the vigorous development of deep learning in recent years, the powerful functions of AI models have been added to the traffic management system to realize intelligent transportation systems (Intelligent Transportation Systems, ITSs). A system that can predict future traffic flow changes with real-time and high accuracy is a very important part of the smart transportation system. The powerful prediction ability of deep learning models enables better results for this short-term traffic flow prediction problem. However, most studies are based on the update of deep learning models, and the pre-processing methods of the dataset and the special characteristics of the dataset are rarely discussed.
Therefore, this paper explored the impact of different data pre-processing methods on traffic flow dataset and the difference between the inter-day feature and intra-day feature of datasets on feature extraction effects of deep learning models. First, many kinds of data denoising methods are used to make datasets more stable and easier to be learned by deep learning models and statistical model. Next, extracts inter- and intra- day traffic flow patterns by a Convolution Neural Network (CNN), then extracted features are imported Long Short-Term Memory (LSTM) units to learn the time series changes of traffic flow. Through a real-data case study of highway situation, extracting intra-day and inter-day features from the data and combine them in different proportions as CNN-LSTM input features has a lower error and better prediction ability than a model that is not suitable for using features and a model that uses single feature in tasks this paper used. This paper shows that the proposed approach can achieve below 5% as the error rate.

1. 大通, David.Luo 時域、頻域、快速傅立葉和同調取樣定理的基本介紹. 2020, 6月, 18日, 擷取自 https://www.wpgdadatong.com/blog/detail/41803
2. 大紀元, "各國人口及汽機車密度比較", 2013年, 05月, 23日, 擷取自 https://reurl.cc/jWQvOM
3. 行政院主計總處, 國情總計通報. 2023, 12月, 26日, 擷取自 https://reurl.cc/Aj60Ad
4. 交通部高速公路局, "國道路網分區圖台北｜ 宜蘭", 擷取自 https://reurl.cc/dn1L0D
5. 交通部高速公路局, "交通資料庫", 擷取自https://tisvcloud.freeway.gov.tw/history-list.php
6. 自由時報, 陳英傑.. 全球每 1000 人的擁車率調查：紐西蘭全球最高 亞洲市場台灣排名第8. 2024, 2月, 擷取自 https://auto.ltn.com.tw/news/25066/3
7. 果仁, "Google 地圖的「即時路況」、「塞車資訊」是怎麼辦到的？", 2019年, 7月, 28日, 擷取自 https://applealmond.com/posts/56198
8. 美國加州聖地牙哥台灣同鄉會, 孫治雄. "淺談信號處理 (A Touch of Signal Processing) ". Oct. 2010. 擷取自https://www.taiwancenter.com/sdtca/articles/10-10/5.html
9. 科技島, "資料科學的2種流派：機器學習與統計模型｜專家論點【維元】", 2022年, 10月, 7日, 擷取自 https://www.technice.com.tw/techmanage/21793/
10. 網路部落格, 從傅立葉級數到快速傅立葉轉換. 2019, 4月, 8日 擷取自 https://blog.yeshuanova.com/2019/04/fft_intro/
11. 網路文章, Fic, "智慧交通", 2021年, 擷取自https://www.fic.com.tw/zh-hant/smart-city/smart-traffic/
12. 網路文章, Hackmd, kuoricky22 Wavelet Transform(小波轉換). 2020, 3月, 19日 擷取自https://hackmd.io/@Youwe/HymbXgZLI
13. 網路文章, Medium, Yasmine Cheng, "模型評估指標：迴歸", 2023年, 3月, 24日, 擷取自 https://reurl.cc/Wx0x8O
14. 網路文章, Vpon, "疫情後觀光趨勢 台灣大數據揭示新興旅遊重心", 2023年, 10月, 18日, 擷取自 https://www.vpon.com/zh-hant/news/202310-taiwan-tourism-data/
15. 網路文章, Yeh James. " [資料分析&機器學習] 第5.1講: 卷積神經網絡介紹(Convolutional Neural Network) ". Dec. 2017. 擷取自https://reurl.cc/NQbqbk
16. A. J. Fowe and Y. Chan, "A microstate spatial-inference model for network-traffic estimation", Transp. Res. C Emerg. Technol., vol. 36, pp. 245-260, Nov. 2013.
17. Aohua Song et.al. “A Hybrid CNN-LSTM Model for Video-Based Teaching Style Evaluation”, in International Conference on Image, Vision and Computing (ICIVC). July. 2023.
18. Alec Radford et al., "Improving Language Understanding by Generative Pre-Training", June. 2018.
19. Anishnama. "Understanding Gated Recurrent Unit (GRU) in Deep Learning". Available online: https://medium.com/@anishnama20/understanding-gated-recurrent-unit-gru-in-deep-learning-2e54923f3e2. May. 2023.
20. B. M. Williams and L. A. Hoel, "Modeling and forecasting vehicular traffic flow as a seasonal ARIMA process: Theoretical basis and empirical results", J. Transp. Eng., vol. 129, no. 6, pp. 664-672, Nov. 2003.
21. B. S., Kerner, S. L. Klenov, G. Hermanns and M. Schreckenberg, "Effect of driver over-acceleration on traffic breakdown in three-phase cellular automaton traffic flow models", Phys. A Stat. Mech. Appl., vol. 392, no. 18, pp. 4083-4105, 2013.
22. Chin Wen Cheong et al., "Asymmetry dynamic volatility forecast evaluations using interday and intraday data", in IEEE Symposium on Business, Engineering and Industrial Applications (ISBEIA), Sep. 2011.
23. Chia-Hsien Cheng, "Efficient VLSI Architecture for Lifting-Based Discrete Wavelet Transform". 2002.
24. Deling Huang., "A Long-term Traffic Flow Prediction Model Based on Multiple Features", in International Conference on Electronic Information Engineering and Computer Communication (EIECC). Dec. 2023.
25. Dai Congcong, Cheng Tiedong, Zong Lu et al., "A study on spectrum characteristics of red sandstone acoustic emission signals based on improved EEMD [J]", Journal of Vibration and Shock, vol. 37, no. 16, pp. 118-123, 2018.
26. Dongfang Ma, Xiang Song & Pu Li, "Daily Traffic Flow Forecasting Through a Contextual Convolutional Recurrent Neural Network Modeling Inter- and Intra-Day Traffic Patterns", in IEEE Transactions on Intelligent Transportation Systems, pp. 2627 - 2636, Feb. 2020.
27. E. I. Vlahogianni, M. G. Karlaftis and J. C. Golias, "Spatio-temporal short-term urban traffic volume forecasting using genetically optimized modular networks", Comput.-Aided Civil Infrastruct. Eng., vol. 22, no. 5, pp. 317-325, 2007.
28. G. Dudek, "Pattern-based local linear regression models for short-term load forecasting", Electr. Power Syst. Res., vol. 130, pp. 139-147, Jan. 2016.
29. Gabriel Mendonça de Paiva et al., "Intelligent Approach to Improve Genetic Programming Based Intra-Day Solar Forecasting Models", in IEEE Congress on Evolutionary Computation (CEC), July. 2018.
30. Hong-Zhe Wei, "Algorithms of Automatic Modulation Classification for Digital/Analog Communication Signals". 2009
31. Hsin Ti Chou, "The analysis and design of amplification circuits biomedical applications". 2012
32. 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 [C]", Proceedings of the Royal Society of London A: Mathematical Physical and Engineering Sciences. Great Britain: The Royal Society, pp. 903-995, 1998.
33. Habte Lejebo Leka et al. "A Hybrid CNN-LSTM Model for Virtual Machine Workload Forecasting in Cloud Data Center". in International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP). 2021.
34. Hanyi Zou, Yuze Wu, Huajun Zhang & Yang Zhan Zhou, "Short-term Traffic Flow Prediction Based on PCC-BiLSTM", in International Conference on Computer Engineering and Application (ICCEA), May. 2020.
35. H. Sun, C. Zhang and B. Ran, "Interval prediction for traffic time series using local linear predictor", Proc. 7th Int. IEEE Conf. Intell. Transp. Syst., pp. 410-415, Nov. 2004.
36. Jia-Wei Kang, "Active Control on the Acoustic Fields in Duct Using the Combined Acoustic Source". June. 2010.
37. Jianwei Liu et al., “Enhanced robustness of myoelectric pattern recognition to across-day variation through invariant feature extraction”, in Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Aug. 2015.
38. K. Cho, B. van Merrienboer, C. Gulcehre, F. Bougares, H. Schwenk, D. Bahdanau, et al., "Learning phrase representations using RNN encoder-decoder for statistical machine translation", arXiv preprint arXiv: 1406.1078, 2014.
39. Liangzhi Zhang et al., "Short-term Traffic Flow Prediction Based on Improved Neural Network with GA", in IEEE International Conference on Artificial Intelligence and Computer Applications (ICAICA). Aug. 2021.
40. Li, Z., Liu, F., Yang, W., Peng, S., & Zhou, J.. "A survey of convolutional neural networks: analysis, applications, and prospects". in IEEE transactions on neural networks and learning systems. 2021.
41. M. S. Ahmed and A. R. Cook, "Analysis of freeway traffic time-series data by using Box-Jenkins techniques", Transp. Res. Rec., no. 722, pp. 1-9, 1979.
42. M. van der Voort, M. Dougherty and S. Watson, "Combining kohonen maps with ARIMA time series models to forecast traffic flow", Transp. Res. C Emerg. Technol., vol. 4, no. 5, pp. 307-318, 1996.
43. M. Cetin and G. Comert, "Short-term traffic flow prediction with regime switching models", Transp. Res. Rec., vol. 1965, no. 1, pp. 23-31, Jan. 2006.
44. M. Cools, E. Moons and G. Wets, "Investigating the variability in daily traffic counts through use of ARIMAX and SARIMAX models: Assessing the effect of holidays on two site locations", Transp. Res. Rec., vol. 2136, no. 1, pp. 57-66, 2009.
45. M. Treiber and A. Kesting, "Validation of traffic flow models with respect to the spatiotemporal evolution of congested traffic patterns", Transp. Res. C Emerg. Technol., vol. 21, no. 1, pp. 31-41, 2012.
46. Marco. "Regarding BiLSTM for timeseries and input/output tensors". Available online: https://discuss.pytorch.org/t/regarding-bilstm-for-timeseries-and-input-output-tensors/173571. Feb. 2023.
47. Ming-Feng Wang, "On the determination of optimal capital structure: Signaling equilibrium and Countersignaling equilibrium". 2004
48. Medium, Rowan. ts, " 9 Techniques for Cross-validating Time Series Data ", Jan 5. 2023. Available online: https://reurl.cc/lQ20jY
49. National Academies, "Transforming Transportation through Connectivity: ITS Strategic Research Plan, 2010–2014", 2012. Available online: https://trid.trb.org/view/1218083
50. Patel, C., Shah, D., & Patel, A., "Automatic number plate recognition system (anpr): A survey". International Journal of Computer Applications, 69(9), 2013.
51. P. Wei, Y. Cao and D. Sun, "Total unimodularity and decomposition method for large-scale air traffic cell transmission model", Transp. Res. B Methodol., vol. 53, pp. 1-16, Jul. 2013.
52. Rui Fu, Zuo Zhang & Li Li , "Using LSTM and GRU neural network methods for traffic flow prediction", in Youth Academic Annual Conference of Chinese Association of Automation (YAC) , Jan, 2016.
53. S. G. X. Xiangjie, "Short-term traffic volume intelligent hybrid forecasting model and its application", Syst. Eng.—Theory Pract., vol. 31, no. 3, pp. 562-568, 2011.
54. S. Lee and D. Fambro, "Application of subset autoregressive integrated moving average model for short-term freeway traffic volume forecasting", Transp. Res. Rec. J. Transp. Res. Board, vol. 1678, no. 1, pp. 179-188, 1999.
55. S. Hochreiter and J. Schmidhuber, "Long short-term memory", Neural Comput., vol. 9, pp. 1735-1780, 1997.
56. Syafrial Fachri Pane et al. "A Hybrid CNN-LSTM Model With Word-Emoji Embedding For Improving The Twitter Sentiment Analysis on Indonesia's PPKM Policy". in International Conference on Information Technology, Information Systems and Electrical Engineering (ICITISEE). Dec. 2022.
57. SingHou Liu, "An Optimal Signal Reconstruction in Noisy Filter Bank System : Multirate Kalman Filtering Approach". 1993
58. Torres M E, Colominas M A, Schlotthauer G et al., "A complete ensemble empirical mode decomposition with adaptive noise [J]", IEEE International Conference on Acoustics Speech and Signal Processing, pp. 4144-4147, 2011.
59. Wu Zhaohua and Norden E. Huang, "Ensemble empirical mode decomposition: a noise-assisted data analysis method [J]", Advances in Adaptive Data Analysis, vol. 1, no. 1, pp. 1-41, 2009.
60. Wenchang Duan et al., "Research and application of urban real-time traffic flow prediction based on STARIMA", in IEEE International Conference on Advances in Electrical Engineering and Computer Applications (AEECA). Aug. 2022.
61. Xiu Zhou et al., "Research on Transformer Partial Discharge UHF Pattern Recognition Based on cnn‐lstm". Dec. 2019.
62. Xinhong You et al., "Optimal day-ahead and intra-day scheduling of energy and operating reserve considering fluctuating wind power", in IEEE International Conference on Environment and Electrical Engineering and IEEE Industrial and Commercial Power Systems Europe (EEEIC / I&CPS Europe), June. 2017.
63. Xia, K., Huang, J., & Wang, "H.. LSTM-CNN architecture for human activity recognition". IEEE Access, 8, 56855-56866. 2020.
64. X. Ma, Z. Tao, Y. Wang, H. Yu and Y. Wang, "Long short-term memory neural network for traffic speed prediction using remote microwave sensor data", Transp. Res. C Emerg. Technol., vol. 54, pp. 187-197, May 2015. 21
65. X.-Y. Xu, J. Liu, H.-Y. Li and J.-Q. Hu, "Analysis of subway station capacity with the use of queueing theory", Transp. Res. C Emerg. Technol., vol. 38, no. 1, pp. 28-43, Jan. 2014.
66. Y. Wang and M. Papageorgiou, "Real-time freeway traffic state estimation based on extended Kalman filter: A general approach", Transp. Res. B Methodol., vol. 39, no. 2, pp. 141-167, 2005.
67. Y. Yuan, J. W. C. Van Lint, R. E. Wilson, F. van Wageningen-Kessels and S. P. Hoogendoorn, "Real-time Lagrangian traffic state estimator for freeways", IEEE Trans. Intell. Transp. Syst., vol. 13, no. 1, pp. 59-70, Mar. 2012.
68. Y. L. Zhang and Y. C. Xie, "Forecasting of short-term freeway volume with V-support vector machines", Transp. Res. Rec., vol. 2024, pp. 92-99, Jan. 2008. 23
69. Y. Lv, Y. Duan, W. Kang, Z. Li and F. Wang, "Traffic Flow Prediction With Big Data: A Deep Learning Approach", IEEE Trans. Intell. Transp. Syst., vol. 16, no. 2, pp. 865-873, Apr. 2015.
70. Y. Tian and L. Pan, "Predicting Short-Term Traffic Flow by Long Short-Term Memory Recurrent Neural Network", in IEEE InternationalConference on Smart City, pp. 153-158, 2015.
71. Yan, S. "Understanding LSTM and Its Diagrams". Available online: https://medium.com/mlreview/understanding-lstm-and-its-diagrams-37e2f46f1714 June. 2018.
72. Yan Cheng Chen, "Identification and Classification of Abnormal Signals for Monitoring or Inspection of Equipment". 2021
73. Z. Shan, D. Zhao and Y. Xia, "Urban road traffic speed estimation for missing probe vehicle data based on multiple linear regression model", Proc. 16th Int. IEEE Conf. Intell. Transp. Syst., pp. 118-123, Oct. 2013.