Estimation method of random number generation method and its parameters follow the multivariate non-normal distribution

摘要

PROBLEM TO BE SOLVED: To develop a distribution to approximate an experiential distribution by a multivariate nonnormal distribution using random numbers, following the Pearson distribution system and the general probability distribution.;SOLUTION: A random number generating method, which generates random numbers, following the multivariate nonnormal distribution based on the Yuan-Bentler method I on a computer, comprises a step of applying n-dimensional multivariate nonnormal distribution to an n-dimensional experiential distribution, by using the computer and a step to generate random numbers using the computer. The applying step performs application about third-order and forth-order moments of experiential distribution, by using the prescribed relational equation of third-order and forth-order moments. Parameters are estimated by a maximum-likelihood method, using the random numbers generating method. The random number generating method and the parameter estimation method are applied to simulations for a financial sector, ion implantation to a semiconductor or the like.;COPYRIGHT: (C)2004,JPO