Variational Inference for Bayesian Networks
[Sourceforge project page]
John Winn, January 2004
Whilst this tutorial used a toy data set, the principles of model construction, modification and comparison can be applied just as readily to real data sets. You can find further examples of networks and both toy and real data sets on the Examples page.
If you intend to use VIBES for real problems, I strongly recommend that you spend some time learning about variational inference and its strengths and limitations. Whilst VIBES is able to make applying variational inference straightforward, unless you have a good understanding of what it is actually doing, you may well encounter problems.
Finally, I would like to emphasise that VIBES is free, unsupported software. I have made every effort to make it bug-free and easy to use, but I cannot provide support. If you find bugs or have feature requests then feel free to log them under Bugs or RFE on the Sourceforge project page, but be aware that there is no guarantee that I will respond to them.