[Forum SIS] seminario B. Huang

[Daniela De Canditiis]

Lunedì 9 Maggio 2016 alle ore 14:00,
presso l’Aula 1B1 dip. SBAI
Università La Sapienza di Roma,
 si terra' il seguente seminario:

SPEAKER: Bill Hsin-Hsiung Huang, Ph.D. (Dep. Of Statistics University of
Central Florida)
TITLE:  An Affine-Invariant Bayesian Cluster Process with Split-Merge Gibbs
Sampler

ABSTRACT:
We develop a clustering algorithm which does not requires knowing the
number of clusters in advance. Furthermore, our clustering method is
rotation-, scale- and translation-invariant coordinatewise. We call it
“Affine-invariant Bayesian (AIB) process”. A highly efficient split-merge
Gibbs sampling algorithm is proposed. Using the Ewens sampling distribution
as prior of the partition and the profile residual likelihoods of the
responses under three different covariance matrix structures, we obtain
inferences in the form of a posterior distribution on partitions. The
proposed split-merge MCMC algorithm successfully and efficiently estimate
the partition. Our experimental results indicate that the AIB process
outperforms other competing methods. In addition, the proposed algorithm is
irreducible and aperiodic, so that the estimate is guaranteed to converge
to the posterior distribution.


Tutti gli interessati sono invitati a partecipare.


Cordiali saluti,
Daniela De Canditiis

-- 
Daniela De Canditiis, PhD
Istituto per le Applicazioni del Calcolo "M.Picone" (CNR)
via dei Taurini, 19 -- 00185 Roma, Italy
tel: +39 06 49270942
fax: +39 06 4404306
http://www.iac.rm.cnr.it/~danielad/
-------------- parte successiva --------------
Un allegato HTML è stato rimosso...
URL: <http://www.stat.unipg.it/pipermail/sis/attachments/20160503/86182802/attachment.html>