I have been reading about semi-supervised techniques and would like to ask if anyone could direct me to any semi-supervised machine learning methods capable of generating probability distribution in a graph-based environment for discrete data? E.g. a Bayesian network like for semi-supervised category.
Tnx for the reference ;)
well, i would like to see the correlation between a huge discrete space (which is Boolean - enable/disable) and correlate them with several other input features (discrete (not bool) / continuous) and performance metrics which will have another distribution itself.
Could you refer me to an implemented package i.e. matlab, c++, etc to be able to test it on my data?
You can also try a semi-supervised graph-label propagation algorithm called Modified Adsorption.
Reference: New Regularized Algorithms for Transductive Learning [ Slides ] [ Video ]
Partha Pratim Talukdar, Koby Crammer, ECML 2009,
Here is a link to the Junto package that implements this algorithm: http://code.google.com/p/junto/
- Badges
- Science topic