BayesFusion Support Forum

For example, when trying to train a network using the PC algorithm on a dataset with both continuous and discrete time series, this is the output that I get in GeNIe: Can't start learning due to problems with data. ## Selected algorithm supports continuous or discrete columns, but they can't be used...

When testing a new data set on a trained network, is it possible to use node states and column values for automatic matching, instead of just labels? In my use case, network nodes and data set columns that correspond to each other have significantly different labels, so the automatic label-based mat...

Will keep an eye out for the package update! Thank you for the explanation :)

Thanks for the prognosis! Since I may not share the actual code, I've tried to recreate the error locally with just 2 files: tutorial1.py and an __init__.py ( pylint_bayesfusion_mwe.zip ). To recreate the error: Create and activate a conda virtual environment with Python 3.12.4 Run: pip install pyli...

As part of my framework's Git CI pipeline, there's a code quality stage that uses Pylint . However, the CI fails with these warnings and error: Problem importing module async.py: No module named '_pysmile' Problem importing module bad_chained_comparison.py: No module named '_pysmile' Problem importi...

So far, I've been working on one machine using one license (academic).

Now I'd like to run some experiments on a cluster — would I need another license for that (possibly also per core), or could I reuse the license from my main machine?

Thank you for the update!

Reducing data requirements for structure learning also sounds like a great way to support more use cases.

In the meantime, I could experiment with filtering those 35k records into a more compact and representative set.

Specifically for PySmile, is it possible to parallelize a function call like pysmile.learning.BayesianSearch.learn(...) ? For example, could we tell SMILE somewhere to use a specific number of threads? Or is there a valid way to divide-and-conquer the learning problem before the "...learn()&quo...

Thanks for the replies and suggestion!

Chance nodes with small bins seems to be working great for now, but I will keep the discretization idea of equation nodes in mind for future reference.

(This post is specifically about the Python wrapper of SMILE) Can prob_evidence(...) be called on a network with equation nodes? For reference, prob_evidence() works for me just fine on networks trained with BS (and CPT nodes). However, after creating a network with PC (using the same training data)...

That would be yet another great feature - thank you!

Thank you for the quick answer!

My bad, I had tried setting the bounds like that last night, but thought I did something wrong when I saw overflow/underflow warning messages. 😅

I see now from this post and the Autodiscretization help page, that this is mostly normal.

I used PC to learn the structure and parameters of a BN (see attached screenshot). The raw data contained time series of 4 temperature sensors in an air handling unit. Now, I'd like to set evidence in the outdoor air temperature (ODA), which intuitively would be a temperature in Celsius degrees betw...

I'd like to learn the parameters of a DBN from the following data set: "Time","Valve setpoint","Valve actual","Outdoor air temperature" 2023-12-07 06:00:00,38.0,38.0,2.0 2023-12-07 08:00:00,35.5,35.5,3.3 2023-12-07 10:00:00,34.0,34.0,5.6 2023-12-07 12:00:00,33...