LAUSR

A copula approach decomposes the joint distribution of random variables into marginal distributions of individual variables and the copula form that links the marginals together. When a researcher is dealing with a modeling problem, he is confronted with obtaining the best possible fit for the observed dependence structure. One possibility is to construct a new ideal copula that can describe the observed dependence. Finding a flexible multi-dimensional copula for modeling dependence is still quite challenging. In this paper, we will construct a new multi-dimensional Archimedean copula function that is characterized by a generator with two parameters which allows for more flexibility in modeling dependence. Moreover, we will apply the new constructed copula on step stress accelerated life testing with dependent competing risks under type II censoring. The point estimates of the unknown parameters are obtained using the maximum likelihood method. Also, the approximate and the parametric bootstrap confidence intervals are constructed. Numerical analysis including simulated data and a real life data about aerospace electrical connector is conducted to study the performance of our proposed copula function.