Hi,
I would greatly appreciate learning of a good paper or manual for
"optim."
More specifically: Often I get error messages and suggestions - as
below "lbfgs" - that are beyond my knowledge and insufficient advice to address
the issue. I have also received "Lapack" messages of the same type.
Thanks,
Paul
Warning message:
In `+.dlm`(mod1, mod2) : the value of mod1$V has been discarded
> fit <- dlmMLE(y, x, expfunc, control = list(trace = 3, REPORT = 2))
N = 19, M = 5 machine precision = 2.22045e-16
This problem is unconstrained.
At X0, 0 variables are exactly at the bounds
At iterate 0 f= 1.8212e+07 |proj g|= 1.2433e+07
At iterate 2 f = 2.3753e+06 |proj g|= 2.8063e+06
At iterate 4 f = 2.9465e+05 |proj g|= 4.7382e+05
At iterate 6 f = 93564 |proj g|= 1.1105e+05
At iterate 8 f = 51371 |proj g|= 26356
At iterate 64 f = 26612 |proj g|= 29.567
ys=-1.949e+01 -gs= 9.047e+00, BFGS update SKIPPED
At iterate 66 f = 26610 |proj g|= 124.86
At iterate 68 f = 26606 |proj g|= 82.252
Bad direction in the line search;
refresh the lbfgs memory and restart the iteration.
Error in optim(parm, logLik, method = method, ...) :
non-finite finite-difference value [3]
In addition: There were 50 or more warnings (use warnings() to see the first 50)
>
[[alternative HTML version deleted]]
_______________________________________________
R-SIG-Mac mailing list
R-SIG-Mac@stat.math.ethz.ch
https://stat.ethz.ch/mailman/listinfo/r-sig-mac
