A New Numerical Algorithm of Three Parameter Generalized Extreme Value Distribution of Wind Speed

Abstract

Maximum likelihood estimation method is used to solve the problem of parameter estimation of three-parameter generalized extreme value distribution. Based on the theory of order reducing，a new numerical algorithm is presented to resolve the problem of maximum likelihood estimation of three-parameter generalized extreme value distribution．Firstly，the shape parameter is assumed to be known and ternary likelihood equations can be transferred into binary ones that are solved with the dichotomy．And then，scale and location parameters are the functions of shape parameter. Further，the maximum likelihood function is described as a unitary function of shape parameter. The optimal estimation of shape parameters can be obtained by applying dichotomy again.