Title :
Unsupervised Learning Approaches for the Finite Mixture Models: EM versus MCMC
Author_Institution :
Sch. of Autom., Hangzhou Dianzi Univ., Hangzhou, China
Abstract :
The key problem in unsupervised learning of finite mixture models is to estimate the parameters of the models. The expectation-maximization (EM) and Markov Chain Monte Carlo (MCMC) are usually used. In this paper, we review these two algorithms and give the complete algorithm processes. We also comment their advantage and disadvantage.
Keywords :
Markov processes; Monte Carlo methods; expectation-maximisation algorithm; unsupervised learning; EM; MCMC; Markov Chain Monte Carlo; expectation maximization; finite mixture models; unsupervised learning approaches; Algorithm design and analysis; Computational modeling; Equations; Estimation; Markov processes; Signal processing algorithms; Unsupervised learning;
Conference_Titel :
Cyber-Enabled Distributed Computing and Knowledge Discovery (CyberC), 2010 International Conference on
Conference_Location :
Huangshan
Print_ISBN :
978-1-4244-8434-8
Electronic_ISBN :
978-0-7695-4235-5
DOI :
10.1109/CyberC.2010.96
