<div dir="ltr"><p class="MsoNormal"><span lang="PT-BR" style="font-size:12pt;line-height:115%;font-family:'Times New Roman',serif">Como pode ser
calculado o intervalo de confiança das estimativas dos parâmetros das EDO (ou modelos não lineares) geradas
pela função modFit, do pacote FME?</span></p>
<p class="MsoNormal"><span lang="PT-BR" style="font-size:12pt;line-height:115%;font-family:'Times New Roman',serif">Abaixo segue o
summay obtido</span></p><p class="MsoNormal"><span lang="PT-BR" style="font-size:12pt;line-height:115%;font-family:'Times New Roman',serif"><br></span></p>
<p class="MsoNormal"><span lang="PT-BR" style="font-family:'Courier New'"> Estimate
Std. Error t value Pr(>|t|) </span></p>
<p class="MsoNormal"><span lang="PT-BR" style="font-family:'Courier New'">Beta1 4.285 0.0041
0.411 0.00687
</span></p>
<p class="MsoNormal"><span lang="PT-BR" style="font-family:'Courier New'">Beta2 0.082 0.0769 0.530 0.00605
</span></p>
<p class="MsoNormal"><span lang="PT-BR" style="font-family:'Courier New'">Beta3 2.041 0.0900 0.411 0.00068</span></p><p class="MsoNormal"><span lang="PT-BR" style="font-family:'Courier New'"><br></span></p>
<p class="MsoNormal" style="text-align:justify"><span lang="PT-BR" style="font-size:12pt;line-height:115%;font-family:'Times New Roman',serif">Eu devo usar a distribuição
t-student retornado pela função, fazendo:</span></p>
<p class="MsoNormal" style="text-align:justify"><span lang="PT-BR" style="font-size:12pt;line-height:115%;font-family:'Times New Roman',serif">Estimate ± Std. Error * t
value</span></p>
<p class="MsoNormal" style="text-align:justify"><span lang="PT-BR" style="font-size:12pt;line-height:115%;font-family:'Times New Roman',serif">Em que </span></p><p class="MsoNormal" style="text-align:justify"><span lang="PT-BR" style="font-size:12pt;line-height:115%;font-family:'Times New Roman',serif"><br></span></p>
<p class="MsoNormal"><span lang="PT-BR" style="font-family:'Courier New'">CI_95%(Beta1)= (4.285</span><span lang="PT-BR">-</span><span lang="PT-BR"> </span><span lang="PT-BR" style="font-family:'Courier New'">0.0041</span><span lang="PT-BR">* </span><span lang="PT-BR" style="font-family:'Courier New'">0.411 ; 4.285</span><span lang="PT-BR">+</span><span lang="PT-BR"> </span><span lang="PT-BR" style="font-family:'Courier New'">0.0041</span><span lang="PT-BR">* </span><span lang="PT-BR" style="font-family:'Courier New'">0.411 )</span><span lang="PT-BR"></span></p>
<p class="MsoNormal"><span lang="PT-BR" style="font-family:'Courier New'">CI_95%(Beta2)= (0.082</span><span lang="PT-BR">-</span><span lang="PT-BR"> </span><span lang="PT-BR" style="font-family:'Courier New'">0.0769</span><span lang="PT-BR">* </span><span lang="PT-BR" style="font-family:'Courier New'">0.530 ; 0.082</span><span lang="PT-BR">+</span><span lang="PT-BR"> </span><span lang="PT-BR" style="font-family:'Courier New'">0.0769</span><span lang="PT-BR">* </span><span lang="PT-BR" style="font-family:'Courier New'">0.530)</span></p>
<p class="MsoNormal"><span lang="PT-BR" style="font-family:'Courier New'">CI_95%(Beta3)= (2.041- 0.0900* 0.411; 2.041+ 0.0900* 0.411)</span></p>
<p class="MsoNormal" style="text-align:justify"><span lang="PT-BR" style="font-size:12pt;line-height:115%;font-family:'Times New Roman',serif"><br></span></p><p class="MsoNormal" style="text-align:justify"><span lang="PT-BR" style="font-size:12pt;line-height:115%;font-family:'Times New Roman',serif">Por que os valores da
distribuição t estão variando, os graus de liberdade não deveriam ser os mesmos
(n-p)? Eem que o n é o tamanho da amostra e p o número de parâmetros?</span></p>
<p class="MsoNormal" style="text-align:justify"><span lang="PT-BR" style="font-size:12pt;line-height:115%;font-family:'Times New Roman',serif">Ou devo usar a distribuição
normal Z(α/2):</span></p><p class="MsoNormal" style="text-align:justify"><span lang="PT-BR" style="font-size:12pt;line-height:115%;font-family:'Times New Roman',serif"><br></span></p>
<p class="MsoNormal"><span lang="PT-BR" style="font-family:'Courier New'">CI_95%(Beta1)= (4.285</span><span lang="PT-BR">-</span><span lang="PT-BR"> </span><span lang="PT-BR" style="font-family:'Courier New'">0.0041</span><span lang="PT-BR">* </span><span lang="PT-BR" style="font-family:'Courier New'">1.96 ; 4.285</span><span lang="PT-BR">+</span><span lang="PT-BR"> </span><span lang="PT-BR" style="font-family:'Courier New'">0.0041</span><span lang="PT-BR">* </span><span lang="PT-BR" style="font-family:'Courier New'">1.96 )</span><span lang="PT-BR"></span></p><p class="MsoNormal"><span lang="PT-BR" style="font-family:'Courier New'"><br></span></p>
<p class="MsoNormal" style="text-align:justify"><span lang="PT-BR" style="font-size:12pt;line-height:115%;font-family:'Times New Roman',serif">Onde eu posso conseguir um
material falando sobre o cálculo dos Intervalos de Confiança produzidos pela
função modFit do pacote FME. O manual do pacote não deixa claro esse cálculo.</span></p>
<p class="MsoNormal" style="text-align:justify"><span lang="PT-BR" style="font-size:12pt;line-height:115%;font-family:'Times New Roman',serif">O cálculo dos intervalos
sempre serão baseados em distribuições simétricas?</span></p><p class="MsoNormal" style="text-align:justify"><span lang="PT-BR" style="font-size:12pt;line-height:115%;font-family:'Times New Roman',serif"><br></span></p><p class="MsoNormal" style="text-align:justify"><span lang="PT-BR" style="font-size:12pt;line-height:115%;font-family:'Times New Roman',serif">Adriele</span></p>-- <br><div class="gmail_signature"><div style="font-family:times,serif;margin:0cm 0cm 0.0001pt;font-size:16px"><span style="white-space:pre"> </span>Adriele Giaretta Biase.</div><div style="font-family:times,serif;margin:0cm 0cm 0.0001pt;font-size:16px"><span style="white-space:pre"> </span>Mestre em Estatística e Experimentação Agropecuária - UFLA. <br><span style="white-space:pre"> </span>Doutoranda em Estatística e Experimentação Agronômica - ESALQ/ USP</div><div style="font-family:times,serif;margin:0cm 0cm 0.0001pt;font-size:16px"><span style="font-size:12pt;white-space:pre"> </span><span style="font-size:12pt">Contato: (19) 8861-0619.</span></div></div>
</div>