Standard error for custom Maximum Likelihood Estimation

2 次查看（过去 30 天）
Marcel
Marcel 2019-3-26
Hey,
i've used the Maximum Likelihood Estimation (https://de.mathworks.com/help/stats/mle.html) with an custom log probability density function to calculate the parameteres of my regression model. Now I'am trying to get the t-statistic of these estimates. I know that I have to divide my estimate by the standard error of the estimate. How do I calculate the standard error?
Currenly I'am using the mlecov funktion (https://de.mathworks.com/help/stats/mlecov.html) but my Estimates seem off.
basismle = mle(data,'logpdf',@custlogpdf,'start',[0,0,0,0,0,0],'Options',options); % custom likelihood estimate
acov = mlecov(phat,data,'logpdf',@custlogpdf) % Asymptotic covariance
se = sqrt(diag(acov)) % Standard Error
tstat = [phat(1,1)/se(1,1); % t-statistic
phat(1,2)/se(2,1);
phat(1,3)/se(3,1);
phat(1,4)/se(4,1);
phat(1,5)/se(5,1);
phat(1,6)/se(6,1)]
Thanks in advance ,
Marcel

请先登录，再回答此问题。

回答（0 个）

类别

在 Help Center 和 File Exchange 中查找有关 Fit Postprocessing 的更多信息

标签

产品


版本

R2018b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by