ABSTRACT
The multivariate parametric estimation has been primarily elaborated for elliptically symmetric distributions, which appeared as an extension of the multi-normal distribution. However, not all symmetric multivariate distribution functions are elliptically symmetric. In the statistical practice and literature, we meet with other norms of the multivariate symmetry, each of them an interest and utilization of their own. Let us briefly describe the concepts of diagonal symmetry, spherical symmetry, total symmetry, marginal symmetry, and symmetry in interchangeability.
