Hi, I currently use OpenCossan, which is free for everyone (academics and industry). Different from NETICA, it is object oriented (it does not yet have a full user interface to support drawing the network), which means you have to understand how to code your network. Once you learn it is very easy and even better. It is free for everyone. The path is: http://www.cossan.co.uk/software/open-cossan-engine.php .

I have also used GeNIe/Smile, recommended by the other colleague, and it is also good - but remember it is only free for academics.

Something important, related to your question: One thing that might be causing your problem: the problem might not be the number of nodes, but the number of links conected to a node (which makes your conditional probability table very large and heavy to compute trhough the standard inference method). Usually the software chooses the exact inference method for you if you don´t select any other. In that case you might try to see if NETICA has the option to choose another inference method that is not exact. I know OpenCossan and Genie give you this option. (For more detail, remember that exact inference algorithms (e.g. computation tree ) are based on analytical approaches, and provide the exact value of the interval probability, while approximation algorithms provide probabilities near the true value. - Which means that if you run the same model a hundred times you will not get exactly the same value if you use an approximation algorithm - which will not be a problem depending on your objective.

I used the HUGIN software for simulating Bayesian networks

You can use statistical software R, and there are several packages for Bayes analysis, and BBN as well. It is free and flexible

Dear Sir, you could use GeNIe academic for Bayesian Network

There are a lot of powerful softwares on the market right now, but they are not necessarily free, for example:

- Hugin Expert

- BayesiaLab

- GeNIe Modeler

You can use a free trial version or ask the developer for a temporary license for academic use.

Hello,

Regarding my experience, I would recommend GeNIe for uncertainty modeling through a Bayesian Network. GeNIe is user-friendly and open-access software.

GeNIe academic