Life Expectancy Estimate With Bivariate Weibull Distribution Using Archimedean Copula

Archimedean copulas are used to construct bivariate Weibull distributions. Co-movementstructures of variables are analyzed through the copulas, where the tail dependence between thevariables is explored with more flexibility. Based on the distance between the copula distributionand its empirical version, a copula that may best fit data is selected. With extra computing costs,the adequacy of the copula chosen is then assessed. When multiple myeloma data areconsidered, it is found that relationship between survival time of a patient and the hemoglobinlevel is well described by the Clayton copula. The bivariate Weibull distribution constructed by thecopula is used to estimate value at risk from which we investigate the anticipated longest lifeexpectancy of a patient with the disease over the treatment period.