A Bayesian Kernel Density Estimator

A Kernel Density Estimator with kernel is defined by

where is the *window width*, *smoothing parameter* or
*bandwidth*. An approach is proposed in which is a parameter
of the problem, so avoiding both the specification of the bandwidth
and the assumption that all projections from a density have the
same smoothness.

- A Bayesian Model
- An Archaeological Problem
- Bayesian Adaptive Kernel Estimates
- Some more testing examples
- Exponential Density
- Maxwell Density
- Cauchy Density
- Infinite Peak Density
- Pareto Density
- Beta (2,2) Density
- Smooth Comb Density
- Sawtooth Density
- Extension to bivariate density estimation
- Discussion

danny 2009-07-23