Generation of a Multivariate Distribution for Specified Univariate Marginals and Covariance Structure.

This paper addresses the problem of determining and generating outcomes for a joint multivariate distribution having specified arbitrary marginal distributions and correlation structure. in addition, a new technique is developed which generates outcome for essentially any given multivariate random variable from a more conveniently realizable random variable. This generalizes in a natural way the well known univariate transformation of probability sampling technique. (Author)