Hi, I have 6 batches of data, each batch contains concentrations of 6 species at 11 time points. My model currently is based on the assumption that the likelihood for each species is normally distributed around the predicted value (the ODE solution) with a different sigma for each species but the s…

The easiest is to reshape the data/prediction to have 3d shape batch x species x time and the you can define sigma per species and broadcast across batch/time. Alternatively you can try to use pt.tile or pt.repeat to replicate along the common dimension

How to set up sigma for multiple batches of data?

Questions version agnostic

ricardoV94 August 30, 2024, 3:36pm 4

Segmentation error is wild. Can you provide a fully reproducible snippet (as simplified as possible) to play with?

Topic		Replies	Views
Shape Parameter for Multiple Treatment Groups Questions	5	654	March 26, 2019
Batched Gaussian Processes Questions	3	572	August 6, 2019
Mu arg in pm.Normal cannot be a vector? Questions	6	1183	October 24, 2019
Using shape parameter vs using specifying it using a different model altogether Questions	5	2023	August 29, 2018
Multilevel model specification Questions	8	1066	January 2, 2020

How to set up sigma for multiple batches of data?

Related topics