Estimating parameters of multivariate regression using the maximum likelihood method with a uniform distribution of residuals?

Representation of regression in a matrix form: $$Y = XA + E,$$

where: $$X$$ - independent variables, $$Y$$ - dependent variables (observations), $$E$$ - errors, which have a uniform distribution, $$A$$ - regression parameter.

How to evaluate the parameters $$A$$ of this multivariate regression using the maximum likelihood method with a uniform distribution of errors?

• Is the range of the uniform distribution given or does it need to be estimated? What do you assume about the independence (or lack thereof) of $E$? – whuber May 27 '19 at 11:45
• @whuber Thank you for the answer! Given: E = [e1, e2, ..., en] ’, // vector of residuals Where ei = U [-ui, ui]; The residuals ei have a uniform distribution with a known range [-ui, ui], All residuals are independent. I need to estimate the vector of parameters A and its variance. – Minh Đại May 27 '19 at 16:52