Construction of a Reactive Computing Environment for Density Estimation in Three to Five Dimensions.

摘要

The focus of this research has been the development of new algorithms and theory for nonparametric estimation of multivariate statistical functions, and the implementation of these algorithms in interesting computing environments. The fundamental algorithm developed for multivariate density estimation has been the so-called averaged shifted histogram. This algorithm has been extended to other statistical functionals, for example, regression and average derivative estimates. Work on visualizing multivariate statistical functions has been presented to a wide audience at scientific meetings and at colloquiums around the country. (kr)