Normally I compute summaries of traces with pm.summary(trace) but this can be quite slow for large problems. Is there an argument to specify that I only want mean and standard deviation? https://docs.pymc.io/api/stats.html?highlight=pymc3%20summary#pymc3.stats.summary specificies that I can pass a …

It looks like you want something like the second example in the documentation, that is, passing a custom stats_funcs dictionary and using extend=False. Example code below: stats_funcs = {"mean": np.mean, "std": np.std} az.summary(idata, stats_funcs=stats_funcs, extend=False, kind="stats") I’d also…

Compute reduced summary statistics

Questions

CorbinFoucart April 30, 2020, 12:56pm 4

Thank you – that’s incredibly helpful. Marking as solution.

1 Like

Topic		Replies	Views
Is there an efficient way to get all of my linear models parameters distributions mean? I.e. analogue to sklearn's .coef_? Questions	5	605	March 26, 2021
Create traces from numpy arrays Questions	7	1398	April 10, 2019
Pm.summary used Quadratic approx? Questions	0	394	May 14, 2019
How to obtain the variance estimations in Linear Regression models Questions	5	889	June 26, 2017
How to print out value of the parameters? Questions development	2	633	April 19, 2018

Compute reduced summary statistics

Related topics