Parallel SMO for Traffic Flow Forecasting

Abstract:

Accurate traffic flow forecasting is crucial to the development of intelligent transportation systems and advanced traveler information systems. Since Support Vector Machine (SVM)have better generalization performance and can guarantee global minima for given training data, it is believed that SVR is an effective method in traffic flow forecasting. But with the sharp increment of traffic data, traditional serial SVM can not meet the real-time requirements of traffic flow forecasting. Parallel processing has been proved to be a good method to reduce training time. In this paper, we adopt a parallel sequential minimal optimization (Parallel SMO) method to train SVM in multiple processors. Our experimental and analytical results demonstrate this model can reduce training time, enhance speed-up ratio and efficiency and better satisfy the real-time demands of traffic flow forecasting.

• Periodical: Applied Mechanics and Materials (Volumes 20-23)
• Pages: 843-848
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.20-23.843
• Online since: January 2010
• Authors: Fan Wang, Guo Zhen Tan, Chao Deng
• Keywords: Parallel SMO, Support Vector Machine (SVM), Traffic Flow Forecasting

References

[1] G.G. He, Y. Li and S.F. Ma: Discussion on Short-Term Traffic Flow Forecasting Methods Based on Mathematical Models. System Engineering Theory Practice, Volume. 12(2000), pp.51-56.

[2] C. Han and S. Song: A review of some main models for traffic flow forecasting. IEEE Intelligent Transportation Systems Proceedings, Volume. 1 (2003), pp.216-219.

DOI: 10.1109/itsc.2003.1251951

[3] V. N. Vapnik: The Nature of Statistical Learning Theory. New York, Springer (1995).

[4] F. Wang, G. Z. Tan, C. Deng, and Z. Tian: Real-time Traffic Flow Forecasting Model and Parameter Selection based on ε-SVR. The 7th World Congress on Intelligent Control and Automation (wcica'08), China, (2008), pp.2870-2875.

DOI: 10.1109/wcica.2008.4593381

[5] C.H. Wu, J.M. Ho, and D.T. Lee: Travel-time prediction with support vector regression. IEEE Transactions on Intelligent Transportation Systems, Vol. 5(4)(2004), pp.276-281.

DOI: 10.1109/tits.2004.837813

[6] Y. Lu and V. Roychowdhury: Parallel Randomized Support Vector Machine. In Proc. PAKDD, pp.205-214, (2006).

[7] Q.S. Yang and C.G. Guo: A Parallel Implementation of Error Correction SVM with Applications to Face Recognition. Lecture Notes in Computer Science, 5552(2009): 327-336.

DOI: 10.1007/978-3-642-01510-6_38

[8] H.P. Graf, E. Cosatto and L. Bottou et . al: Parallel support vector machines: the cascade SVM [ C ] Advances in Neural Information Processing Systems, Cambridge, MA: M IT Press, pp.521-528 (2005).

[9] J. Yang. An imp roved cascade SVM training algorithm with crossed feedbacks[C]. IEEE Proceedings of IMSCCS'06 , Hangzhou, Vol. 2(2006), pp.735-738.

DOI: 10.1109/imsccs.2006.183

[10] L. J. Cao, S. S. Keerthi, C. J. Ong, J. Q. Zhang, U. Periyathamby, X.J. Fu, and H. P. Lee, Parallel sequential minimal optimization for the training of support vector machines. IEEE Trans. Neural Network, vol. 17(2006), no. 4, pp.1039-1049.

DOI: 10.1109/tnn.2006.875989

[11] J.C. Platt. Fast Training of support Vector Machines Using Sequential Minimal Optimization[M]. Cambridge, MA: MIT Press, pp.185-208(1999).

DOI: 10.7551/mitpress/1130.003.0016

[12] N.Y. Deng and Y.J. Tian: New Method of Data Mining-Support Vector Machines [M] . Beijing : Science Press , pp.224-274(2004).

[13] Freeway Performance Measurement System (PEMS), version7. 0 [Online]: http: /pems. eecs. berkeley. edu/Public.