Research

next up previous
Next: Classification Up: Support independent statistics Previous: Estimation of support size

Estimating information theoretic quantities

Application areas: classification, biology

While estimation of entropy or mutual information is not identical to estimation of the underlying distribution, a good approach to the distribution estimation problem does help, and is currently the best approach known under some conditions [21]. Results from [13] show promise in using a plug-in estimator based on distribution modeling from patterns.



Prasad Santhanam 2007-12-28