<div dir="ltr"><p class="MsoNormal" style="text-align:justify;text-justify:inter-ideograph"><b><span style="font-size:10pt;font-family:Arial,sans-serif;background-image:initial;background-repeat:initial">Seminário - Projeto Temático: Modelos de
Regressão e Aplicações</span></b></p>

<p class="MsoNormal" style="text-align:justify;text-justify:inter-ideograph"><b><span style="font-size:10pt;font-family:Arial,sans-serif;background-image:initial;background-repeat:initial"> </span></b></p>

<p class="MsoNormal" style="text-align:justify;text-justify:inter-ideograph"><b><span lang="EN-US" style="font-size:10pt;font-family:Arial,sans-serif;background-image:initial;background-repeat:initial">Título: </span></b><span lang="EN-US" style="font-size:9.5pt;font-family:Arial,sans-serif;background-image:initial;background-repeat:initial">Multivariate Measurement
Error Models Based on Student-t Distribution Under Censored Responses</span><b><span lang="EN-US" style="font-size:10pt;font-family:Arial,sans-serif;background-image:initial;background-repeat:initial"></span></b></p>

<p class="MsoNormal" style="text-align:justify;text-justify:inter-ideograph"><b><span lang="EN-US" style="font-size:10pt;font-family:Arial,sans-serif;background-image:initial;background-repeat:initial"> </span></b></p>

<p class="MsoNormal" style="text-align:justify;text-justify:inter-ideograph"><b><span style="font-size:10pt;font-family:Arial,sans-serif;background-image:initial;background-repeat:initial">Palestrante: </span></b><span style="font-size:10pt;font-family:Arial,sans-serif;background-image:initial;background-repeat:initial">Celso R. B. Cabral, Depto. de Estatística, Universidade Federal do Amazonas</span></p>

<p class="MsoNormal" style="text-align:justify;text-justify:inter-ideograph"><b><span style="font-size:10pt;font-family:Arial,sans-serif;background-image:initial;background-repeat:initial"> </span></b></p>

<p class="MsoNormal" style="text-align:justify;text-justify:inter-ideograph"><b><span style="font-size:10pt;font-family:Arial,sans-serif;background-image:initial;background-repeat:initial">Quando: </span></b><span style="font-size:10pt;font-family:Arial,sans-serif;background-image:initial;background-repeat:initial">21 de agosto, sexta-feira, às 11h.</span></p>

<p class="MsoNormal" style="text-align:justify;text-justify:inter-ideograph"><span style="font-size:10pt;font-family:Arial,sans-serif;background-image:initial;background-repeat:initial"> </span></p>

<p class="MsoNormal" style="text-align:justify;text-justify:inter-ideograph"><b><span style="font-size:10pt;font-family:Arial,sans-serif;background-image:initial;background-repeat:initial">Onde: </span></b><span style="font-size:10pt;font-family:Arial,sans-serif;background-image:initial;background-repeat:initial">Auditório Jacy Monteiro, Bloco B, piso térreo - IME-USP</span></p>

<p class="MsoNormal"><span style="font-size:9.5pt;font-family:"Arial",sans-serif"> </span></p>

<p class="MsoNormal" style="text-align:justify;background-image:initial;background-repeat:initial"><b><span lang="EN-US" style="font-size:9.5pt;font-family:"Arial",sans-serif">Resumo</span></b><span lang="EN-US" style="font-size:9.5pt;font-family:"Arial",sans-serif">. </span><span lang="EN-US" style="font-size:9.5pt;font-family:Arial,sans-serif">Measurement
error models constitute a wide class of models, that include linear and
nonlinear regression models, very useful to model many real life phenomena, in
particular in the medical and biological areas. The great advantage of this
type of models is that, in some sense, they can be represented as mixed effects
models, allowing to us the implementation of well-known techniques, as the
EM-algorithm for the parameter estimation. In this work, we consider the study
of a class of multivariate measurement error models where the observed response
and/or the covariate is not fully observed, or in other words, the observations
are subjected to certain threshold values below or above which the measurements
are not quantifiable. Consequently, these observations are considered as
censored. We assume a Student-t distribution as the distribution of the
unobserved true value of the covariate and the error term of the model,
providing a robust alternative to parameter estimation. Our approach relies on
a likelihood-based inference using the EM-algorithm. The proposed method is
illustrated through a simulation study and a real data set.</span></p>

<p class="MsoNormal" style="background-image:initial;background-repeat:initial"><span lang="EN-US" style="font-size:9.5pt;font-family:Arial,sans-serif"> </span></p>

<p class="MsoNormal" style="background-image:initial;background-repeat:initial"><span lang="IT" style="font-size:9.5pt;font-family:Arial,sans-serif">Joint work with Larissa A. Matos (UNICAMP), Luis
M. Castro (Universidad de Concepción) and Víctor H. Lachos (UNICAMP)</span></p></div>