Error message after increasing the number of iterations about bayesiantools

[florianhartig]

Hi yihsiu,

thanks for reporting this. It's hard to say what the problem is without seeing your code. Would you mind posting your code + data here, or if you feel uncomfortable with this, send it to me via email?

Best,
F

[yihsiu]

Hi Florian,

Thanks very much for reply. Since it's a long code, it'd be more readable in the email. I will send you an email in a wee while.

Thanks again!

Best wishes
Yi-Hsiu

[florianhartig]

The issue was that, in the script, the adaptation start was set before the end of the burnin.

The way the sampler is currently programmed, this is not possible, because the adaptation is based on the recorded chain. I have now added checks so that the user is warned if such a settings is attempted in 20f489c

Yi-Hsiu, I've now also rerun the sampler, once this is removed all seems to work fine.

A general note: the burnin setting is intended for saving memory. If you don't have a memory problem, better not touch it, there is no harm in recording the whole chain. All later functions in BayesianTools have the option start = XX to discard the desired amount of burnin from the calculations / plots.