MATLAB: Does ‘fitglm’ use least squares or maximum likelihood to form the fit

glmleast squaresMATLABmaximum likelihood

When I run test code, it appears to me that 'fitglm' is using least squares and NOT maximum likelihood. Can you confirm?

Best Answer

The 'fitglm' function makes use of the “Iteratively Reweighted Least Squares” method to find the regression coefficients in the maximum likelihood stage of the function.

The MATLAB documentation link of 'fitglm' lists a few reference books which have more information regarding this:

http://www.mathworks.com/help/stats/fitglm.html

Related Solutions

MATLAB: Log likelihood

If you have the most recent release of the Statistics Toolbox:

>> x = poissrnd(4,20,1);
>> pd = fitdist(x,'poisson');
>> pd.NLogL
ans =
   39.0221

If you do not:

>> mu = poissfit(x);
>> -sum(log(poisspdf(x,mu)))
ans =
   39.0221

MATLAB: Quantifying fit of data to distribution

You can do it visually like this: documentation. Alternatively, try fitmethis on FileExchange. I think it computes the likelihood of the data under each fitted distribution and indicates the one under which the data have the highest likelihood.

Best Answer

Related Solutions

MATLAB: Log likelihood

MATLAB: Quantifying fit of data to distribution

Related Question