[R] fgev_error_matrix_singular

[Jinrui Xu]

Hello everyone,

I am using fgev from evd package to fitting my data with general  
extreme value distribution. The command I used is simple: y <- fgev(x).
The error is "observed information matrix is singular;" How can I  
solve these problem. I tried all methods in optim and it did not work.  
I fixed values for scale, location and shape. This works, but I have  
lots of different dataset to fit, so is there any automatic ways to  
solve the error? Thanks!

jinrui,