Hi all. I’m trying to predict on new data, but pm.sample_posterior_predicitve does not catch the new shape of data. I have tried an approach combining this and this posts, but to no avail. Here’s my latest attempt: import numpy as np import pymc as pm rng = np.random.default_rng(18) score = np.ra…

Your predictions don’t look right. In a time series model you should never have wiggles and dynamics in the average prediction unless you built them in, i.e. with seasonal components or AR terms. For a GRW, your best guess of tomorrow’s value is today’s value, so your predictions should be a straigh…

Sample_posterior_predicitve not catching shape of new data

Questions v5

cluhmann August 18, 2022, 8:59pm 5

Try this?

[Edit:] Which clearly won’t work. I would keep an eye on that issue.

1 Like

Topic		Replies	Views
Out of sample prediction v5 modeling	20	1964	April 7, 2023
Sample_ppc shape Questions	5	2688	July 18, 2018
How to predict new values on hold-out data Questions	24	13526	July 22, 2020
Understanding shape of values returned by sample_posterior_predictive Questions	3	922	January 26, 2020
Using pm.Data to predict on two inputs for sample_posterior_predictive; why is there no change in the results? Questions	4	1183	May 17, 2021

Sample_posterior_predicitve not catching shape of new data

Related topics