p*s
发帖数: 71 | 1 I used R-package JohnsonDistribution to estimate Sb curve parameters using
method of moments, but R shows no solution. Then, I tried SAS and Fortran.
SAS also can not find a solution, but Fortran did. The Fortran program was
published on a journal. I believe the R-package and SAS proc come from the
Fortran program. I wonder why Fortran can obtain results but SAS and R
cannot do it. | p*s
发帖数: 71 | 2 I think, a number generated during estimation is too close to zero that SAS
and R regarded it as zero or too large to be out of range allowed by SAS and
R because in Fortran results a parameter estimate is very large. I am not
sure if it is the reason. Since SAS and R show no solution, I wonder if I
can believe Fortran results. |
|
