The obvious: all optimization procedures can be subject to such errors, and simplifying models and/or weakening the criteria for stopping (increasing step size, increasing the delta of the change in value that signals convergence, trying different starting values for the parameters, etc.) are more or less standard procedures to get convergence. It can be tricky -- numerical optimization is part art as well as science.
If you don't get a satisfactory reply here or at SO, you might wish to contact the package maintainer: see ?maintainer. Cheers, Bert On Sun, Jan 29, 2023 at 2:53 PM ZAC WARHAM <zac.war...@students.mq.edu.au> wrote: > Hi all, I am receiving the error *Newton failed to find minimum* whilst > trying to fit a move persistence model in this package. I think the error > is related to the optimiser in the TMB package but it is beyond my > statistics/coding knowledge to track down the specific cause and solution. > I would appreciate it if anyone knew some resources to assist with fixing > this. I have put a MRE and sample data over on Stackoverflow if anyone > wants a closer look - > > https://stackoverflow.com/questions/75253642/how-do-i-fix-newton-failed-to-find-minimum-in-r > > -- > > [[alternative HTML version deleted]] > > ______________________________________________ > R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide > http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. > [[alternative HTML version deleted]] ______________________________________________ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
