Multivariate Poisson regression

#1
I have a linear model y=Ax where y and x are vectors and A is a matrix. The dimensionality of y is more than that of x. Further y has Poissonian noise, so more precisely y = Poissonian(Ax)

Given the observed y vector and known A matrix, I need to find the MLE of the x vector.

I need help locating software that can do this. So far all the software I've found can do this only for single response variable y. They don't work for vector y. These include Matlab's glmfit, SPSS' GENLIN, SAS GENMOD, STATA poisson and R glm.

Thanks!
 
Last edited:

Link

Ninja say what!?!
#2
Why don't you just write out the equation for the MLE and have a calculating software (i.e. excel) do it for you?

I have a linear model y=Ax where y and x are vectors and A is a matrix. The dimensionality of y is more than that of x. Further y has Poissonian noise, so more precisely y = Poissonian(Ax)

Given the observed y vector and known A matrix, I need to find the MLE of the x vector.

I need help locating software that can do this. So far all the software I've found can do this only for single response variable y. They don't work for vector y. These include Matlab's glmfit, SPSS' GENLIN, SAS GENMOD, STATA poisson and R glm.

Thanks!
 
#3
Why don't you just write out the equation for the MLE and have a calculating software (i.e. excel) do it for you?
I'll implement it in Matlab myself if I have to. It's a very large data set and I would much rather use code that's already been optimized though.