What are the limitations on the data file when I want to learn a Bayesian Network's structure? I have a file with about 208 thousand lines and 30 variables in each line, and when I try to learn the structure GeNIe crashes.
If GeNIe/SMILE can't run with datasets as large as this, does anyone know of any other similar tools that could handle it?
Limitations on data file size when learning BN structure?
-
pedromvilar
- Posts: 1
- Joined: Wed Aug 26, 2015 9:11 am
-
marek [BayesFusion]
- Site Admin
- Posts: 430
- Joined: Tue Dec 11, 2007 4:24 pm
Re: Limitations on data file size when learning BN structure?
GeNIe does not impose any limitations, so the real limitation is your available memory. If the file is too large for your computer, perhaps you can try selecting a random sample from among the file records? A random sample should retain the file's property and should lead to a model that will reflect the joint probability distribution that has generate the file.
Good luck!
Marek
Good luck!
Marek