Multi-thread search with deterministic annealing EM algorithm

Author

Takada, Masaharu ; Nakano, Ryohei

Author_Institution

Nagoya Inst. of Technol., Japan

Volume

1

fYear

2002

fDate

6/24/1905 12:00:00 AM

Firstpage

1034

Lastpage

1038

Abstract

The EM algorithm is an efficient algorithm to obtain the ML estimate for incomplete data, but has the local optimality problem. The deterministic annealing expectation-maximization (DAEM) algorithm was once proposed to solve the problem, but is not guaranteed to obtain the global optimum since it employs a single token search. The paper investigates the possibility of the multiple-thread search with the DAEM algorithm for a Gaussian mixture. The experiments showed the minimal beam size to guarantee the global optimality is not so large for a Gaussian mixture, and the solution quality of the beam DAEM algorithm always exceeds the EM and DAEM algorithms

Keywords

Gaussian distribution; multi-threading; search problems; simulated annealing; Gaussian mixture; ML estimate; deterministic annealing EM algorithm; efficient algorithm; global optimum; incomplete data; local optimality problem; minimal beam size; multithread search; Annealing; Bifurcation; Clustering algorithms; Computational efficiency; Cooling; Embedded computing; Iterative algorithms; Maximum likelihood estimation; Temperature;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks, 2002. IJCNN '02. Proceedings of the 2002 International Joint Conference on

Conference_Location

Honolulu, HI

ISSN

1098-7576

Print_ISBN

0-7803-7278-6

Type

conf

DOI

10.1109/IJCNN.2002.1005618

Filename

1005618