Bayesian network for continuous variables

Question

I have search and saw some questions on the matter but without answer (due to the fact that the questions were asked more than 1 year ago, I. hoped something has changed)

I am looking for a library to infer bayesian network from a file of continious variables is there anything simple\out of the box that any one has encountered? I have tried pyAgrum for example but when i run

pyAgrum.BNLearner(numdata).learnDAG()

I get

Exception: [pyAgrum] Wrong type: Counts cannot be performed on continuous variables. Unfortunately the following variable is continuous: V0

Have tried serval libraries but they all seem to work only on discrete variables would love some help in advance.

Answer 1

The main question is what kind of model do you want for your continuous variables.

1- Do you want them to be discretized : you can have a look for instance at http://webia.lip6.fr/~phw/aGrUM/docs/last/notebooks/Discretizer.ipynb.html.

2- Do you want to assume a linear gaussian model : you can have a look for instance at bnlearn (https://haipengu.github.io/Rmd/GBN.html)

3- Do you want to learn more general continuous model : You can have a look at for instance otagrum (http://openturns.github.io/otagrum/master/) which learns copula bayesian network.

4- etc.