<div dir="auto"><div>Divulgando.<div class="gmail_quote"><br><br><div dir="ltr"> <span></span> <p class="m_-2537962864183530920gmail-western" align="justify" style="margin-bottom:0cm;line-height:100%;font-family:Calibri,sans-serif;font-size:12pt;direction:ltr;color:rgb(0,0,0)"> <strong><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><span style="background:rgb(255,255,255)">Seminário - Modelos de Regressão e Aplicações</span></font></font></font></strong><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><br> <br> </font></font></font><strong><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><span style="background:rgb(255,255,255)">Título</span></font></font></font></strong><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><span style="background:rgb(255,255,255)">:</span></font></font></font><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><span style="background:rgb(255,255,255)"> </span></font></font></font><span style="font-variant:normal"><font color="#222222"><font face="arial, sans-serif"><font size="2" style="font-size:9pt"><span style="letter-spacing:normal"><span style="font-style:normal"><span style="font-weight:normal"><span style="background:rgb(255,255,255)">Invariant prior distributions for hierarchical models</span></span></span></span></font></font></font></span></p><p class="m_-2537962864183530920gmail-western" align="justify" style="margin-bottom:0cm;line-height:100%;font-family:Calibri,sans-serif;font-size:12pt;direction:ltr;color:rgb(0,0,0)"><strong><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><span style="background:rgb(255,255,255)">Palestrante</span></font></font></font></strong><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><span style="background:rgb(255,255,255)">: Thais Fonseca – </span></font></font></font><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><span style="background:rgb(255,255,255)">DME -</span></font></font></font><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><span style="background:rgb(255,255,255)"> UFRJ</span></font></font></font></p><p class="m_-2537962864183530920gmail-western" align="justify" style="margin-bottom:0cm;line-height:100%;font-family:Calibri,sans-serif;font-size:12pt;direction:ltr;color:rgb(0,0,0)"><strong><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><span style="background:rgb(255,255,255)">Quando</span></font></font></font></strong><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><span style="background:rgb(255,255,255)">: </span></font></font></font><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><span style="background:rgb(255,255,255)">03</span></font></font></font><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><span style="background:rgb(255,255,255)"> de </span></font></font></font><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><span style="background:rgb(255,255,255)">agosto</span></font></font></font><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><span style="background:rgb(255,255,255)"> de 2018, sexta-feira, às 1</span></font></font></font><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><span style="background:rgb(255,255,255)">4</span></font></font></font><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><span style="background:rgb(255,255,255)">h.</span></font></font></font><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><br> <br> </font></font></font><strong><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><span style="background:rgb(255,255,255)">Onde</span></font></font></font></strong><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><span style="background:rgb(255,255,255)">:</span></font></font></font><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><span style="background:rgb(255,255,255)"> </span></font></font></font><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt">Sala 2</font></font></font><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt">43</font></font></font><font color="#222222"><font face="Arial, sans-serif"><span style="background:rgb(255,255,255)"> </span></font></font><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt">– Bloco A, segundo andar - IME-USP<br> <br> </font></font></font><b>Resumo</b><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt">.</font></font></font><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><span style="background:rgb(255,255,255)"> </span></font></font></font><span style="font-variant:normal"><font color="#222222"><font face="Arial, sans-serif"><font size="2" style="font-size:9pt"><span style="letter-spacing:normal"><span style="font-style:normal"><span style="font-weight:normal"><span style="background:rgb(255,255,255)">In this work, we propose an approach for constructing invariant prior distributions for hierarchical models based on information geometry. In particular, we provide a flexible decomposition for the Fisher information tailored for multilevel models. This proposal takes advantage of the hierarchy and leads to an alternative way of computing Jeffreys priors for the hyperparameters besides providing an upper bound for the prior information. While the Jeffreys prior gives the minimum information about parameters, the proposed bound gives an upper limit for the information put in the an objective prior. Any prior with information above that limit may be considered too informative. Furthermore, the proposed prior may be computed computationally as part of a MCMC algorithm. This property might be essential in models with many levels in which analytical marginalization is not feasible. As an alternative, we suggest an objective prior which may be proper when the usual Jeffreys prior is not. We illustrate the usefulness of our proposal with examples in mixture models, in selection priors such as lasso and in the student-t model.</span></span></span></span></font></font></font></span></p> <br></div> </div></div></div>