An approach based on a global optimization method is developed in this paper for efficient estimation of parameters of the Pearson type 3 distribution. A global optimization method called Alienor is coupled with a stochastic search technique. This approach replaces the method of maximum likelihood estimation where it failed (a case where the shape parameter b is less than 1) and also in other cases because it provides better simulation results. To experiment our new approach we use the R statistical software and we compare our approach to the classical methods.

Alienor method, a-dense, O.P.O, estimation, likelihood, Pearson type 3 model, global optimization.