Dec. 20 2005 9:00AM

Dec. 20 2005 9:30AM

Density estimation plays a key role in a broad variety of decision-making processes. Recent developments have introduced new kind of estimates with nice theoretical behavior and computationally efficient. We review some of them together with their properties with respect to information criteria. Combinatorial methods seem promising to build these estimates automatically from data. We show how they can be used to select parameters or to partition the observation space.

International Workshop on Data - Algorithms - Decision Making (Kontaktní osoba: Martin Janžura)