[Forum SIS] Seminario ALESSIO FARCOMENI (Univ. di Roma La Sapienza) - Padova, 26 novembre 2009

[Patrizia Piacentini]

Ci scusiamo per trasmissioni multiple
***********************************
DIPARTIMENTO DI SCIENZE STATISTICHE
Universita' degli Studi di Padova
Via C. Battisti, 241 - 35121 Padova
***********************************

Seminario

A CLASSICAL AND A BAYESIAN SOLUTION TO THE LIKELIHOOD
DILEMMA IN RECAPTURE STUDIES WITH HETEROGENEITY

tenuto da

ALESSIO FARCOMENI
Università di Roma "La Sapienza"

Giovedì 26 novembre 2009
ore 14.30
Aula Cucconi


Abstract
Capture-recapture experiments are commonly used to estimate the size of 
a closed population. Link (2003) shows that inference is not possible 
without assumptions on the mixing distribution of heterogenous capture 
probabilities, due to identifiability problems. We show that 
identifiability does not hold only for the commonly used conditional 
likelihood. We prove the marginal likelihood is instead identifiable and 
the MLE consistent. We illustrate different strategies for obtaining the 
MLE with the marginal likelihood. We also propose a Bayesian approach, 
derive its Jeffreys prior, and compare the two methods in simulations 
and on real examples.

http://www.stat.unipd.it/agenda/seminari

Tutti gli interessati sono invitati a partecipare

-- 
Patrizia Piacentini
Segreteria Organizzativa Eventi
tel  049 8274167 - fax 049 8274170
e-mail segrorg at stat.unipd.it