[ABE-L] Ciclo de Palestras - PPGE-IM/UFRJ - 10 de junho

[Carlos Zanini]

 Caros alunos e professores,

A próxima palestra do Ciclo de Palestras do Programa de Pós-graduação em
Estatística (PPGE) da UFRJ ocorrerá *quarta-feira, 10 de junho, *no horário
das *15h00* às 16h30. A palestra será *transmitida remotamente *a partir do
Laboratório de Sistemas Estocásticos (LSE), Sala I-044-B, Centro de
Tecnologia - UFRJ, com acesso remoto através do link:
meet.google.com/hno-grzn-ujb

*Palestrante*: Daiane Aparecida Zuanetti

*Title*: Applying LASSO idea to select variables in non-homogeneous hidden
Markov models

*Abstract*: Non-homogeneous hidden Markov models are a statistical paradigm
in which a sequence of non-observable states generates a sequence of
observations. One of their great uses has been to understand, describe and
predict environmetric patterns (such as rain, pollution, some environmental
disasters, etc) of specific areas of the world as functions of other
covariates. Transitions between the non-observable states (different
periods) are controlled by transition coefficients and covariates. Since
variable selection has been hardly explored for this model, the central
purpose of this study is to propose scalable variable selection methods
which improve predictive performance of the model. We propose two versions
of the LASSO for the non-homogeneous hidden Markov model. The proposals
consistently show better predictive performance than ARIMA and Penalized
Linear Regression and very good performance when predicting the
non-observable state sequence which generates the observable values. In
terms of coefficient shrinkage efficiency, the proposals show excellent
performance in all simulation scenarios when selecting variables via
coefficient shrinkage. Finally, the methods are applied to characterize and
predict the rainfall regime in the city of São Carlos, Brazil, which is our
main motivation. This is a joint work with Gustavo Alexis SAbillón. G. A.
Sabillón was partially supported by CAPES and D. A. Zuanetti was supported
in part by FAPESP (process number: 2024/00413-8).

Att
Carlos.



-- 
----------------------------------------------------------------------------------------
Carlos Tadeu Pagani Zanini
Professor Adjunto
Departamento de Métodos Estatísticos - UFRJ
SIAPE: 1146101
carloszanini em dme.ufrj.br
carloszanini em im.ufrj.br
https://sites.google.com/view/ctpzanini
-------------------------------------------------------------------------------------------
-------------- Próxima Parte ----------
Um anexo em HTML foi limpo...
URL: <http://lists.ime.usp.br/pipermail/abe/attachments/20260608/a04f5aee/attachment.htm>