Advances and Applications in Statistics
Volume 48, Issue 6, Pages 445 - 471
(June 2016) http://dx.doi.org/10.17654/AS048060445 |
|
SAMPLING TARGET DISTRIBUTIONS BASED ON SUPPORT ALTERATION AND DISCRETIZATION
Saleem Z. Ramadan
|
Abstract: We propose a non-iterative, non-parametric and restriction-free sampling method based on altering the support of the target distribution from an unequally likely support to an equally likely support. The comparative analysis carried out shows that this method is applicable for various distributions whether they are univariate or multivariate and whether they have a continuous or a discrete and a bounded or an unbounded support. In addition, the comparative analysis shows that this method provided competitive results with respect to the existing methods including slice sampling. Based on the method’s computational times, we believe that the computational time is satisfactory. |
Keywords and phrases: sampling methods, discretization, multidimensional density functions, equally likely support, support alteration. |
|
Number of Downloads: 436 | Number of Views: 1454 |
|