GeNIe crash, too much variables?

[EllenMeijerink]

Hey!

I am working with GeNIe for my graduation thesis. I am creating a bayesian network to analyze the impacts of various variables on 'working from home' and how working from home influences other activities.
However, when I upload my data file, create background knowledge and select the variables that should be included, the program seems to crash and closes off. If I add less variables, the network is created. I am working on a 64GB desktop. Does GeNie need more than 64GB? Or what else can be the problem here?
Thank you!

Kind regards,

Ellen

[shooltz[BayesFusion]]

GeNIe Academic is a 32-bit program, and has access to 2 gigabytes of user address space, regardless of the installed RAM. You can try the free 30-day trial version of business version of GeNIe, available for download at:
https://download.bayesfusion.com/files. ... y=Business
GeNIe for business is available in both 32-bit and 64-bit versions.

[marek [BayesFusion]]

Hi Ellen,

If the algorithm that you are using is the PC algorithm, it is possible that the structure that it discovers is too complex. There may be nodes that have many parents. In that case, the crash occurs when you go to the next step, i.e., learn parameters. The probability tables are huge and may lead to exhausting your available memory. You could do several things to avoid this: (1) lower the Significance Threshold to very small, which will generally lead to a sparser graph, (2) limit the number of parents (Max Parent Count), (3) impose a time limit in case the algorithm takes a long time. Finally, you can also try another algorithm, such as Bayesian Search or Greedy Thick Thinning. I would be happy to help/suggest something if I had the data set.
Cheers,

Marek

[EllenMeijerink]

Hey Marek,

Thank you very much!
How can I lower the Significance Threshold?
And how could I use Bayesian Search or Greedy Thick Thinning?

I am willing to share the dataset, how can I do that?

Thank you!

[marek [BayesFusion]]

Both parameters are accessible from the Learn New Network dialog. The upper right pop-up menu allows you for choosing the algorithm. Have you actually looked at GeNIe manual :-)? You can send me the data set by a private message on the Forum or by Email (marek@bayesfusion.com).
Cheers,

Marek