LargeScale Traffic Congestion Prediction Based on the Symmetric Extreme Learning Machine Cluster Fast Learning Method
Abstract
:1. Introduction
2. Methodology
2.1. Extreme Learning Machine
2.2. Symmetric Extreme Learning Machine Algorithm
2.3. Proposed Symmetric Extreme Learning Machine Cluster Algorithm
Algorithm 1. SELMCluster fast learning 
Train Input: Training sample X, sample marker T, number of hidden layer neurons L, regularization coefficient C, excitation function $\tilde{g}\left(x\right)$. Output: All enumerated types $\left\{{v}_{1},{v}_{2},\dots ,{v}_{K}\right\}$, hidden layer output weights group $\beta =\left\{{\beta}_{1},{\beta}_{2},\dots ,{\beta}_{K}\right\}$, shared hidden layer network $\tilde{h}\left(x\right)$.

3. ShortTerm Congestion Prediction Based on the Symmetric Extreme Learning Machine Cluster Algorithm
3.1. Time Series Analysis of Traffic Flow
3.2. Traffic Congestion Index
3.3. Feature Extraction and Modelling
3.3.1. Section Clustering
3.3.2. Feature Extraction
 Road factors: These include the road level, lane number, road traffic light distribution, road intersection distribution, and road related inherent features.
 Environmental factors: These include the urban functional areas on which roads are located, and whether there are schools, large public places or other factors that affect the traffic flow.
 Sudden factors: These include uncertain factors such as weather changes, traffic accidents, traffic control, and other key activities.
 Road clusters, discrete features, 1, 2, 3, 4, …, 50, a total of 50 kinds of values.
 The current time, discrete characteristics, 06:05, 06:15, 21:55, a total of 191 values.
 The congestion values in the past eight historical periods: A continuous feature with a range of 0–100, wherein each historical period is ten minutes.
 Road level: Highways, expressways, main roads, secondary roads and branches.
 The number of adjacent roads at the road entrance: This continuous feature is 0, 1, 2, 3.
 The number of adjacent road connections at the road section: Continuous characteristics, with a value of 0, 1, 2, 3, ….
4. Implementation and Experimental Results
4.1. Data
4.2. SELMCluster Model Tuning Test
4.2.1. The Length of Time Series
4.2.2. The Regularization Coefficient C and the Number of Hidden Layer Nodes L
4.3. Comparison Test of the SELMCluster and ELM
4.4. Comparison of the SELMCluster and Other Algorithms
4.5. Prediction Results Evaluation
5. Conclusions
Author Contributions
Acknowledgments
Conflicts of Interest
References
 Ghosh, B.; Basu, B.; O’Mahony, M. Bayesian TimeSeries Model for ShortTerm Traffic Flow Forecasting. J. Transp. Eng. 2007, 133, 180–189. [Google Scholar] [CrossRef]
 Guo, J.; Huang, W.; Williams, B.M. Adaptive Kalman filter approach for stochastic shortterm traffic flow rate prediction and uncertainty quantification. Transp. Res. Part C 2014, 43, 50–64. [Google Scholar] [CrossRef]
 Yu, G.; Zhang, C. Switching ARIMA model based forecasting for traffic flow. In Proceedings of the International Conference on Acoustics, Speech, and Signal Processing, Brighton, UK, 12–17 May 2004; Volume 2, pp. 429–432. [Google Scholar]
 Zhang, Y.; Ye, Z. ShortTerm Traffic Flow Forecasting Using Fuzzy Logic System Methods. J. Intell. Transp. Syst. 2008, 12, 102–112. [Google Scholar] [CrossRef]
 Xue, J.; Shi, Z. ShortTime Traffic Flow Prediction Based on Chaos Time Series Theory. J. Transp. Syst. Eng. Inf. Technol. 2008, 8, 68–72. [Google Scholar] [CrossRef]
 Cetin, M.; Comert, G. ShortTerm Traffic Flow Prediction with Regime Switching Models. J. Transp. Res. Board 2006, 1965, 23–31. [Google Scholar] [CrossRef]
 Abdulhai, B.; Porwal, H.; Recker, W. ShortTerm Traffic Flow Prediction Using NeuroGenetic Algorithms. ITS J. Intell. Transp. Syst. 2002, 7, 3–41. [Google Scholar] [CrossRef]
 Smith, B.L.; Demetsky, M.J. ShortTerm Traffic Flow Prediction: Neural Network Approach Transportation Research Record. Transp. Res. Board 1994, 98–104. [Google Scholar]
 Zhao, L.; Wang, F.Y. Shortterm traffic flow prediction based on ratiomedian lengths of intervals twofactors highorder fuzzy time series. In Proceedings of the Vehicular Electronics and Safety, Beijing, China, 13–15 December 2007; pp. 1–7. [Google Scholar]
 Xu, Q.; Ding, Z. Bi discipline. A realtime prediction method of traffic flow based on dynamic recurrent neural network. J. Huaihai Inst. Technol. Nat. Sci. Edit. 2004, 12, 14–17. [Google Scholar]
 Yang, Q.; Zhang, B.; Gao, P. Based on improved dynamic recurrent neural network for short time prediction of traffic volume. J. Jilin Univ. Eng. Edit. 2012, 4, 887–891. [Google Scholar]
 Jiao, G.T.; Xu, J.; Ma, Y. Traffic volume prediction method based on QPSORBF. Traffic Inf. Saf. 2008, 26, 128–131. [Google Scholar]
 Çetiner, B.G.; Sari, M.; Borat, O.A. Neural Network Based TrafficFlow Prediction Model. Math. Comput. Appl. 2010, 15, 269–278. [Google Scholar] [CrossRef]
 Mussone, L. A review of feedforward neural networks in transportation research. E I Elektrotech. Informationstech. 1999, 116, 360–365. [Google Scholar] [CrossRef]
 Bing, Q.; Gong, B.; Yang, Z.; Shang, Q.; Zhou, X. A shortterm traffic flow local prediction method of combined kernel function relevance vector machine. J. Harbin Inst. Technol. 2017, 49, 144–149. [Google Scholar]
 Shang, Q.; Lin, C.; Yang, Z.; Bing, Q.; Zhou, X. Shortterm traffic flow prediction model using particle swarm optimization–based combined kernel functionleast squares support vector machine combined with chaos theory. Adv. Mech. Eng. 2016, 8, 1–12. [Google Scholar] [CrossRef]
 Huang, G.B.; Zhu, Q.Y.; Siew, C.K. Extreme Learning Machine: Theory and Applications. Neurocomputing 2006, 70, 489–501. [Google Scholar] [CrossRef]
 Aguirre, L.; Lopes, R.; Amaral, G.; Letellier, C. Constraining the topology of neural networks to ensure dynamics with symmetry properties. Phys. Rev. E 2004, 69, 1–11. [Google Scholar] [CrossRef] [PubMed]
 Chen, S.; Wolfgang, A.; Harris, C.; Hanzo, L. Symmetric RBF classifier for nonlinear detection in multipleantennaaided systems. Trans. Neural Netw. 2008, 19, 737–745. [Google Scholar] [CrossRef] [PubMed]
 Chen, S.; Hong, X.; Harris, C. Greybox radial basis function modelling. Neurocomputing 2011, 74, 1564–1571. [Google Scholar] [CrossRef]
 Espinoza, M.; Suykens, J.; Moor, B. Imposing symmetry in least squares support vector machines regression. In Proceedings of the 44th IEEE conference on decision and control, Seville, Spain, 12–15 December 2005; pp. 5716–5721. [Google Scholar]
 McNames, J.; Suykens, J.A.K.; Vandewalle, J. Winning entry of the K. U. Leuven time series prediction competition. Int. J. Bifurc. Chaos 1999, 9, 1485–1500. [Google Scholar] [CrossRef]
 Liu, X.; Li, P.; Gao, C. Symmetric extreme learning machine. Neural Comput. Appl. 2013, 22, 551–558. [Google Scholar] [CrossRef]
 Box, G.E.P.; Jenkins, G.M.; Reinsel, G.C.; Ljung, G.M. Time Series Analysis: Forecasting and Control, 5th Edition. J. Oper. Res. Soc. 2015, 22, 199–201. [Google Scholar]
 Shumway, R.H.; Stoffer, D.S. Time Series Analysis and Its Applications: With R Examples; Springer: Berlin, Germany, 2017. [Google Scholar]
 Akaike, H. A new look at the statistical model identification. IEEE Trans. Autom. Control 1974, 19, 716–723. [Google Scholar] [CrossRef]
 Anděl, J.; Perez, M.G.; Negrao, A.I. Estimating the dimension of a linear model. Kybern.Praha 1981, 17, 514–525. [Google Scholar]
Unblocked  Basic Unblocked  Mild Congestion  Moderate Congestion  Serious Congestion  

Speed interval (km/h)  Highways, expressways  (65,∞)  (50,65]  (35,50]  (20,35]  [0,20] 
Main road  (40,∞)  (30,40]  (20,30]  (15,20]  [0,15]  
Secondary roads, branches  (35,∞)  (25,35]  (15,25]  (10,15]  [0,10]  
Congestion value  (0,20)  [20,40)  [40,60)  [60,80)  [80,100]  
Map show color  Light green  Green  Yellow  Red  Deep red 
Highways, Expressways  Main Roads  Secondary Roads, Branches 

0.028  0.052  0.065 
Training Data Set  Training Sample Source  Number of Training Samples  Test Sample Source  Number of Test Samples 

Working day  20–24 March 2017  4,961,698  27 March 2017  1,036,785 
Weekend  Saturdays and Sundays, March 2017  5,023,698  26 March 2017  968,254 
Major festival  Ching Ming Festival and May Day in 2017  5,069,547  2 May 2017  1,187,234 
Algorithm  Training Time (s)  Accuracy 

ELM  116  64.28% 
SELM  67  65.77% 
SELMCluster (single process)  134  92.99% 
SELMCluster (20 processes)  9.2  93.12% 
SELMCluster (40 processes)  5.1  93.04% 
Algorithm  Training Time (s)  Accuracy 

SELMCluster (40 processes)  24.63  92.08% 
GBDT  4501.16  92.10% 
Ridge  893.23  89.49% 
Lasso  9.34  86.79% 
LR  7.69  74.15% 
Algorithm  Training Time (s)  Accuracy 

SELMCluster (single process)  10.57  90.91% 
SVM  686.85  87.87% 
© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Share and Cite
Xing, Y.; Ban, X.; Liu, X.; Shen, Q. LargeScale Traffic Congestion Prediction Based on the Symmetric Extreme Learning Machine Cluster Fast Learning Method. Symmetry 2019, 11, 730. https://doi.org/10.3390/sym11060730
Xing Y, Ban X, Liu X, Shen Q. LargeScale Traffic Congestion Prediction Based on the Symmetric Extreme Learning Machine Cluster Fast Learning Method. Symmetry. 2019; 11(6):730. https://doi.org/10.3390/sym11060730
Chicago/Turabian StyleXing, Yiming, Xiaojuan Ban, Xu Liu, and Qing Shen. 2019. "LargeScale Traffic Congestion Prediction Based on the Symmetric Extreme Learning Machine Cluster Fast Learning Method" Symmetry 11, no. 6: 730. https://doi.org/10.3390/sym11060730