Date: Oct 28, 2013 6:50 AM
Author: Alessandro Beretta
Subject: MLE biivariate multiple regression
I'm trying to develop a function in order to estimate the parameter of a bivariate multiple regression via maximum-likelihood estimation method.
The model is as follows:
Y = X*b + a
where Y and a are nx2 matrices, X is a nx5 matrix and b is 5x2 matrix with the coefficients that I need to estimate (n is the number of observation in the dataset).
What I need is a function in order to minimize the negative log likelihood function:
sum(for i=1:n) of [0.5*log|cov| + 0.5*(Yi-Xi*b)'*cov^-1*(Yi-Xi*b)]
I know that I'll need also 1 function for the optimization constraints and 1 function for the variance covariance matrix (cov) estimation.
Does anybody could help me coding this function???
Thanks a lot!