Weird mode check plots

Estimation ML/Bayesian estimation
1

Dear sir:

I am implementing a Bayesian estimation, and I get weird mode check plots: mode_check_1.pdf (8.6 KB) .

Although the the estimated mode of parameter “A” is at the maximum of the posterior likelihood, but the log-post curve is not smooth.

Does this estimation find the proper mode? Thank you very much.

Best Regards

Paul

2

Look at the scaling of the x- and y-axes. The mode is plotted around a small region and the differences in the posterior value are tiny. This typically comes from a parameter running to a prior bound. This usually is a warning sign.

3

Dear professor Pfeifer:

Thank you very much.

Best Regards

Paul