Calculate median from Gaussian Inference

Questions
1

Hello, I’m a novice in the Bayesian field and have been reading the great introductory book by Osvaldo Martin. Having said that, I’m have made some inferences on my blood pressure with data recorded from my blood pressure monitor. The cut off value for high blood pressure is 140/90 mmHg. I’m worried about my systolic pressure since the 95% HDI includes the cut off value of 140 mmHg, although there’s a 0.90 probability that the real mean is below that threshold. Does that mean I have high systolic blood pressure? How could I best explain these results to my physician?

Also, why does the Arviz summary table doesn’t show values for the median and the mode of the posterior distribution? How can I infer these parameters if possible?

One last question, how can make a 2D contour plot with the posterior distributions of my systolic and diastolic pressure means?

Thank you.

presionsys

2

Hi Felipe,

az.summary displays the mean by default, but you can use the stat_funcs kwarg to add the median to the summary, as shown in an example in the doc.

Not sure which plot you’re looking for, but I’d advise looking at ArviZ API – there is probably what you’re after.
Hope this helps :vulcan_salute:

1 Like
3

Thank you very much!

1 Like
4

Regarding the 2D contour plot, I think what you want is the plot_kde function:
https://arviz-devs.github.io/arviz/examples/matplotlib/mpl_plot_kde_2d.html

2 Likes