Hello, I am trying to replicate the below piece of code (which is in pymc3) to the latest pymc version: with model_2: prior_pred = pm.sample_prior_predictive() fig, ax = plt.subplots(figsize = (20, 8)) _ = ax.plot(prior_pred["likelihood"].T, color = "0.5", alpha = 0.1) _ = ax.plot(data_tr…

It might be good to check the xarray documentation . If you just want a mean over chains and draws, then you would do idata.prior_predictive.likelihood.mean(dim=['chain', 'draw']). You don’t need to sel in this case, since both dims will be reduced by the mean. This discards all the distributional i…

Plotting the prior likelihood vs target in pymc

Questions v5

jessegrabowski February 17, 2024, 11:56am 2

pm.sample_prior_predictive adds a dummy chain dimension (that’s the 1 at the end of the shape there). You can get rid of it with .sel(chain=0), or by squeezing the array that .values returns.

Topic		Replies	Views
Sample_prior_predictive bug, gives wrong shape for dependent variable? Questions	2	458	July 29, 2019
Prior predicitive for linear regression Questions	2	544	May 22, 2019
Plot posterior predictive several coordinates Questions	7	1728	September 24, 2020
Sample_prior_predictive failing with shape argument Questions	7	506	May 10, 2021
ValueError: x and y can be no greater than 2D, but have shapes (144,) and (144, 500, 144) Questions	4	3646	May 24, 2022

Plotting the prior likelihood vs target in pymc

Related topics