Estimation Procedure for a New Bivariate Distribution

We introduced a new bivariate distribution for a random vector Z = [X, Y]^T, where X and Y are positive continuous variables. The distribution is a generalization of the Weibull distribution, it has 7 parameters: r and s are power parameters, m and n are moment parameters, a and b are scaling parameters, and p is the linking parameter. The ML estimation turns out to be a very difficult task. In this paper we present the estimation procedure. It was tested on simulated data, which we generated using the acceptance-rejection method. The results are very satisfactory.