I am looking for solutions similar to leaky Noisy-OR gates in Bayesian networks, that allow for handling uncertainty caused by small amount of training data (see attached publication).
The solution I am looking for does not have to refer to Bayesian networks.
I would appreciate some literature references, or just a key-words so I know what to search for.
Article Learning Bayesian Network Parameters from Small Data Sets: A...
I am working on an evolution of the LAD approach (Logical Analysis of Data, Boros Hammer et al.) with the precise aim of learning from small training sets. If you have datasets, you can contact me at [email protected]
Why not to try a preprocessing phase by bootstrap?
http://aclweb.org/anthology//E/E03/E03-1008.pdf
After that it'll be possible applying other appropriate methods.
Depends on your final goal !
For classification purpose let have a look at TANB,
else, for very small dataset (or with high missing data rate), the minimum weighting spanning for tree-shaped-Bayesian (MWST) network could be a starting point if you wanna stay with BN formalism !
Yes it depends on the goal as said Olivier. If a probabilistic model is interesting to reach your objectives you can see the papers :
Simon C., Weber P., Evidential networks for reliability analysis and performance evaluation of systems with imprecise knowledge. IEEE Transactions on Reliability, 58, 1, March (2009), 69-87. DOI : 10.1109/TR.2008.2011868
Simon C., Weber P., Imprecise reliability by evidential networks. Proceedings of the Institution of Mechanical Engineers, Part O: Journal of Risk and Reliability 223, 2, (2009), 119-131. DOI : 10.1243/1748006XJRR190
And we have working on leaky Noisy-OR gates formulation in evidential networks:
Fallet G., Weber P., Simon C. Iung B., Duval C., Evidential network-based extension of Leaky Noisy-OR structure for supporting risks analyses. In 8th International Symposium SAFEPROCESS 2012, Mexico, august (2012).
you can contact me form more information...
- Badges
- Science topic