Dear allstaters
When trying to estimate 3 positive parameters of a continuous distribution
by the method of moments
(by equating the first 3 population central moments with their sample
counterparts)
I obtain negative values for one or two parameters!
P.S. The method of maximum likelihood is very difficult for my
distribution, since
the p.d.f. includes a special function called confluent hypergeometric
function
in 2 of the parameters!
Any suggestions!
Thanks in advance
M. Ghitany
|