0

I am new with Bayesian Inference in Python and I wish to perform Bayesian updating with new data every time it is generated (10-min frequency). This means I will not perform it on a known set of data but update the model with singular datum every time. Unfortunately, I have not found any way of doing so with Python yet. Does anybody have a suggestion on how to proceed in such case? I was trying to use PYMC3 to set my model up. I will be very grateful for any help. Kind regards

mats_snaps
  • 57
  • 6
  • Does this answer your question? [Incremental model update with PyMC3](https://stackoverflow.com/questions/40870840/incremental-model-update-with-pymc3) Also, [this other question](https://stackoverflow.com/questions/53211277/updating-model-on-pymc3-with-new-observed-data) has a particular implementation that takes advantage of a conjugacy. – merv Mar 21 '20 at 16:45
  • Thanks for this suggestion. It seems that in both of those cases, the users suggest solutions to quite complex problems. I was hoping to find a way to build the simplest possible model that would update itself every timestep. I am analysing a wind farm and its' data comes evevry 10 mins. – mats_snaps Mar 22 '20 at 12:27
  • Consider adding an example of your current code to fit a single set of data. Without a proposed model, it's difficult to suggest an online updating strategy. – merv Mar 22 '20 at 19:54

0 Answers0