Hello,dear Gretl Community.
I'm doing a nonlinear estimation via the maximum likelihood method. I want
to do a batch processing of data but in some cases the estimation don't
satisfise tolerance (1,81899e-012), don't met the convergence criterion.
Does anyone knows how it is calculated whether or not the model meets
An estimation that don't met tolerance;
[image: Imágenes integradas 1]
An estimation that met tolerance;
[image: Imágenes integradas 2]
[image: Imágenes integradas 3]
Suggestion: I think it's calculated by adding up logslikelihood or
gradients. But I do both and don't coincided with tolerance.
Another question, How can I save automatically results?
Beforehand, thank you for any help.
Mario Alberto Flórez Porras
Student of Industrial Engineering
I'm interested to know if there are any possibilities to generate panel
plots from a script, or to retrieve the mean of the selected series.
Also the panel plot command doesn't appear in the command log.
#there are some missing observation
smpl YEAR>1978 --restrict
smpl YEAR<1983 --restrict
#no missing observation
# for every industry in which the company operates do a panel plot for wage
smpl IND=$i --restrict
summary WAGE --simple
# select WAGE, right click select panel plot/group means
#is it posibile to generate a panel plot in a loop
# or at least to retrive the information wich is used ( the mean of the
series for every year???
From: Mario Florez Porras <mario.florez.porras(a)gmail.com>
Sent: Tuesday, August 30, 2016 4:36 PM
Subject: [Gretl-users] MLE Advanced
Anyone know how it is calculated the covariance matrix?
Here is Octave code which produces the same covariance values as gretl:
degrees_of_freedom = cases-predictorscoefficients = pinv(IV)*DV
estimates = IV*coefficients;
residuals = DV-estimates;
SSres = residuals'*residuals;
residual_variance = SSres/degrees_of_freedom; N=IV'*IV;
covariance_matrix = Ni*residual_variance;