METHOD OF CLUSTERIZATION, BASED ON SERIAL LAUNCHING OF K-MEANS WITH IMPROVED SELECTION OF THE CANDIDATE FOR NEW POSITION OF INSERTION
Keywords:
code books, clusterization, k-means, centroides, kd-treesAbstract
The paper suggests improved method of K-means clusterization, which, unlike the conventional method, allows to obtain the solution, close to global minimum of distortion by means of serial launching of k-means for 1, 2, ..., k .centroides. Decrease of distortion is achieved at the expense of improvement of the procedure of vectors-candidates definition on the selection of insertion position of new centroide without considerable slowing down of operation time.
REFERENCES
1. Fayyad U. M. Advances in Knowledge Discovery and Data Mining / U. M. Fayyad, G. Piatetsky-Shapiro, P. Smyth, R. Uthurusamy // AAAI/MIT Press. – 1996. – 611 p.
2. Gersho A. Vector Quantization and Signal Compression. / A. Gersho, R. M. Gray // Boston: Kluwer Academic. – 1992. – 760 p.
3. Duda R. O. Pattern Classification and Scene Analysis / R. O. Duda, P. E. Hart // New York: John Wiley & Sons. – 1973. – 512 p.
4. Jain A. K. Algorithms for Clustering Data / A. K. Jain, R. C. Dubes // Englewood Cliffs, N.J.: Prentice Hall. – 1988. – 334 p.
5. Arthur D. k-means++: The advantages of careful seeding / D. Arthur, S. Vassilvitskii // ACM-SIAM Symposium on Discrete Algorithms (SODA 2007) Astor Crowne Plaza – New Orleans, Louisiana. – 2007. – P. 1027 – 1035.
6. Lai Jim Z. C. Fast k-means clustering algorithm using cluster center displacement / Jim Z. C. Lai, Tsung-Jen Huang, Yi-Ching Liaw // Pattern Recognition. – 2009. – No 11, vol. 42. – P. 2551 – 2556.
7. Kanungo T. An Efficient k-means clustering algorithm: analysis and implementation / T. Kanungo, D. M. Mount, N. S. Netanyahu, C. Piatko, R. Silverman and A.Y. Wu // IEEE Transactions On Pattern Analysis And Machine Intelligence. – 2002. – No. 7, vol. 24. – P. 881 – 892.
8. Likas A. The global k-means clustering algorithm / Aristidis Likas, Nikos Vlassis, Jacob J. Verbeek // Pattern Recognition. – 2002. – No 2, vol. 36. – P. 451 – 461.
9. Refining initial points for KMeans clustering : (Conference on Machine Learning) [Електронний ресурс] / P. S. Bradley, U. M. Fayyad // Proceedings of Fifteenth Intl. – 1998. – P. 91 – 99. / Режим доступу: ftp://ftp.research.microsoft.com/pub/tr/tr-98-36.pdf.
10. Hussein N. A Fast Greedy K-Means Algorithm / Master’s Thesis Nr:9668098 N. Hussein. – University of Amsterdam Faculty of Mathematics, Computer Sciences, Physics and Astronomy Euclides Building Plantage muidergracht 24. – 2002. – p. 62.
11. Moore Andrew William Efficient memory based learning for robot control / Moore Andrew William. – PhD thesis Nr: UCAM-CL-TR-209. – 1990. – p. 248.
12. Kanungo T. A Local Search Approximation Algorithm for k-Means Clustering / T. Kanungo, D. M. Mount, N. S. Netanyahu, C. Piatko, R. Silverman, A. Y. Wu // Computational Geometry: Theory and Applications. – 2004. – No 2. – P. 89 – 112.
13. Ткаченко О. М. Ефективне векторне квантування LSF-параметрів при ущільненні мовних сигналів / О. М. Ткаченко, О. Д. Феферман, С. В. Хрущак // Інформаційні технології та комп’ютерна інженерія. – 2007. – № 1. – С. 124 – 129
Downloads
-
PDF
Downloads: 140
