 |
Center for Nonlinear Studies |
In the area of Bayesian non-parametric methods, Polya trees provide a method for creating flexible probability models with absolutely continuous distribution functions. Their use has historically been limited by the curse of dimensionality. This talk will present work in progress on a new simplex-based scheme for partitioning multidimensional space that drastically reduces the cost of increasing dimensionality.
Host: Emily Michele Casleton