Thank you very much for the help and the great suggestion on cleaning up the
residual function.
Setting the bounds and being very careful with the starting position (on my
real data) do seem to do the trick, which I wouldn't have guessed given the
error message that was getting thrown.
Best,
Did you notice that the residual function I gave flipped what your
original example problem used as starting values for munb and size? so you
can sometimes afford to give a bad starting value for these param.
in trying to calculate the gradient L-BFGS-B may try evaluate fn with
parameter values
2 matches
Mail list logo
