An Algorithm for High-Dimensional Traffic Data Clustering

  • Pengjun Zheng
  • Mike McDonald
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4223)


High-dimensional fuzzy clustering may converge to a local optimum that is significantly inferior to the global optimal partition. In this paper, a two-stage fuzzy clustering method is proposed. In the first stage, clustering is applied on the compact data that is obtained by dimensionality reduction from the full-dimensional data. The optimal partition identified from the compact data is then used as the initial partition in the second stage clustering based on full-dimensional data, thus effectively reduces the possibility of local optimum. It is found that the proposed two-stage clustering method can generally avoid local optimum without computation overhead. The proposed method has been applied to identify optimal day groups for traffic profiling using operational traffic data. The identified day groups are found to be intuitively reasonable and meaningful.


Discrete Fourier Transform Traffic Data Optimal Partition Initial Partition Partition Matrix 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Jain, A.K., Dubes, R.C.: Algorithms for Clustering Data. Prentice-Hall, Englewood Cliffs (1988)MATHGoogle Scholar
  2. 2.
    Jain, A.K., Murty, M.N., Flynn, P.J.: Data Clustering: A Review. ACM Computing Surveys 31(3), 264–323 (1999)CrossRefGoogle Scholar
  3. 3.
    Hand, D.J., Krzanowski, W.J.: Optimising k-means Clustering Results with Standard Software Packages. Computational Statistics & Data Analysis 49(4), 969–973 (2005)MATHCrossRefMathSciNetGoogle Scholar
  4. 4.
    Dunn, J.C.: A fuzzy relative of the ISODATA process and its use in detecting compact well separated clusters. Journal of Cybernetics 3, 32–57 (1974)CrossRefGoogle Scholar
  5. 5.
    Bezdek, J.C.: Pattern Recognition with Fuzzy Objective Function Algorithms. Plenum Press, New York (1981)MATHGoogle Scholar
  6. 6.
    Selim, S.Z., Ismail, M.A.: K-means Type Algorithms: A Generalised Convergence Theorem and Characterisation of Local Optimality. IEEE Transactions on Pattern Analysis and Machine Intelligence 6(1), 81–87 (1984)MATHCrossRefGoogle Scholar
  7. 7.
    Pollard, D.: A Central Limit Theorem for k-Means Algorithm. Annals of Probability 10, 919–926 (1982)MATHCrossRefMathSciNetGoogle Scholar
  8. 8.
    Murthy, C.A., Chowdhury, N.: In Search of Optimal Clusters using Genetic Algorithms. Pattern Recognition Letters 17, 825–832 (1996)CrossRefGoogle Scholar
  9. 9.
    Jones, D., Beltramo, M.A.: Solving Partitioning Problems with Genetic Algorithms. In: Proceedings of Fourth International Conference of Genetic Algorithms, pp. 442–449 (1991)Google Scholar
  10. 10.
    Laszlo, M., Mukherjee, S.: A Genetic Algorithm Using Hyper-Quadtrees for Lowdimensional K-means clustering. IEEE Transactions on Pattern Analysis and Machine Intelligence 28(4), 533–543 (2006)CrossRefGoogle Scholar
  11. 11.
    Vanloan, C.F.: Generalizing Singular Value Decomposition. SIAM Journal on Numerical Analysis 13(1), 76–83 (1976)CrossRefMathSciNetGoogle Scholar
  12. 12.
    Keogh, E.J., Pazzani, M.J.: A simple dimensionality reduction technique for fast similarity search in large time series databases. In: Terano, T., Chen, A.L.P. (eds.) PAKDD 2000. LNCS, vol. 1805, pp. 122–133. Springer, Heidelberg (2000)CrossRefGoogle Scholar
  13. 13.
    Jagadish, H.V., Koudas, N., Muthukrishnan, S., Poosala, V., Sevcik, K., Suel, T.: Optimal Histograms with Quality Guarantees. In: Proceedings of the 24th VLDB Conference, New York, USA (1998)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2006

Authors and Affiliations

  • Pengjun Zheng
    • 1
  • Mike McDonald
    • 1
  1. 1.Transportation Research GroupUniversity of SouthamptonHighfield, SouthamptonUnited Kingdom

Personalised recommendations