IEICE Technical Committee Submission System Conference Paper's Information Online Proceedings [Sign in] ... (for ESS/CS/ES/ISS) Tech. Rep. Archives ... (for ES/CS)
 Go Top Page Go Previous [Japanese] / [English]

 Paper Abstract and Keywords Presentation 2010-11-05 15:30 [Poster Presentation] Efficient Computation of Normalized Maximum Likelihood Coding for Gaussian Mixtures with Its Applications to Model SelectionSo Hirai, Kenji Yamanishi (Tokyo Univ.) Abstract (in Japanese) (See Japanese page) (in English) We are concerned with the issue of efficient computation of normalized maximum likelihood (NML) code-lengths for Gaussian mixture distributions with its applications to model selection.Although the NML code-length is the optimal one in that it achieves Shtarkov's minimax criterion,there is a problem that the normalization factor in the NML code-length becomes infinity in the case where data is continuous.In this paper we propose a method for computation of the NML code-length for Gaussian mixture distributions so that the normization term is bounded by a finite value and the total computation time is ${\rm O}(n^2)$($n$ is sample size).We empirically demonstrate the effectiveness of the proposed method in terms of the rate of convergence of the selected number of clusters in comparison with AIC and BIC. Keyword (in Japanese) (See Japanese page) (in English) Normalized Maximum Likelihood / Minimum Description Length Principle / Gaussian Mixture Distribution / / / / / Reference Info. IEICE Tech. Rep., vol. 110, no. 265, IBISML2010-103, pp. 327-333, Nov. 2010. Paper # IBISML2010-103 Date of Issue 2010-10-28 (IBISML) ISSN Print edition: ISSN 0913-5685  Online edition: ISSN 2432-6380

 Conference Information Committee IBISML Conference Date 2010-11-04 - 2010-11-06 Place (in Japanese) (See Japanese page) Place (in English) IIS, Univ. of Tokyo Topics (in Japanese) (See Japanese page) Topics (in English) IBIS 2010 (Workshop on Information-based Induction Sciences) Paper Information Registration To IBISML Conference Code 2010-11-IBISML Language Japanese Title (in Japanese) (See Japanese page) Sub Title (in Japanese) (See Japanese page) Title (in English) Efficient Computation of Normalized Maximum Likelihood Coding for Gaussian Mixtures with Its Applications to Model Selection Sub Title (in English) Keyword(1) Normalized Maximum Likelihood Keyword(2) Minimum Description Length Principle Keyword(3) Gaussian Mixture Distribution Keyword(4) Keyword(5) Keyword(6) Keyword(7) Keyword(8) 1st Author's Name So Hirai 1st Author's Affiliation The University of Tokyo (Tokyo Univ.) 2nd Author's Name Kenji Yamanishi 2nd Author's Affiliation The University of Tokyo (Tokyo Univ.) 3rd Author's Name 3rd Author's Affiliation () 4th Author's Name 4th Author's Affiliation () 5th Author's Name 5th Author's Affiliation () 6th Author's Name 6th Author's Affiliation () 7th Author's Name 7th Author's Affiliation () 8th Author's Name 8th Author's Affiliation () 9th Author's Name 9th Author's Affiliation () 10th Author's Name 10th Author's Affiliation () 11th Author's Name 11th Author's Affiliation () 12th Author's Name 12th Author's Affiliation () 13th Author's Name 13th Author's Affiliation () 14th Author's Name 14th Author's Affiliation () 15th Author's Name 15th Author's Affiliation () 16th Author's Name 16th Author's Affiliation () 17th Author's Name 17th Author's Affiliation () 18th Author's Name 18th Author's Affiliation () Speaker 1 Date Time 2010-11-05 15:30:00 Presentation Time 180 Registration for IBISML Paper # IEICE-IBISML2010-103 Volume (vol) IEICE-110 Number (no) no.265 Page pp.327-333 #Pages IEICE-7 Date of Issue IEICE-IBISML-2010-10-28