I think the imputation feature in PyMC3 is fantastic and really streamlines some workflows. However, I have noticed in some applications that it leads to slower NUTS sampling. For example, if I construct a regression model with N fully observed data points, it will sample faster than the same model …

I see. This is in general difficult to deal with. Either you fatten your prediction and observed, or use pm.Potential with mask like what you did in TFP.

Disabling missing data imputation

Questions

ricardoV94 August 25, 2023, 6:18am 9

You can also just try pm.logp(rv[mask], value[mask]) inside a Potential.

1 Like

Topic		Replies	Views
Missing values in a model? Questions	12	4791	November 7, 2018
Automatic imputation and posterior predictive sampling with new data Questions	5	2068	October 29, 2020
Masking missing values of predictors Questions	3	1385	July 10, 2020
Dealing with missing data and custom distribution Questions	13	2248	March 14, 2021
Calculating the likelihood based on not missing observed values Questions	5	1995	June 12, 2018

Disabling missing data imputation

Related topics