This is the directory of the matlab functions that I have
written. The following zip files can be expanded into
subdirectories of functions:
utility - a miscellaneous collection of functions (THESE FILES ARE USED
BY .M FILES IN EBBS, LPOLY, AND penalized_splines)
ebbs - local bandwidth selection for local polynomial regression
lpoly - local polynomial regression
penalized_splines - penalized regression splines
The functions in ebbs, lpoly, and srs call functions in utility.
Also, functions in ebbs call functions in lpoly.
Here's some documentation for EBBS. srs contains a README file
with documentation. I haven't yet developed
documentation for the other files.
EBBS
This is a collection of functions for local polynomial
smoothing with a local bandwidth computed by EBBS (Ruppert,
1997, JASA). The .m files in this directory use .m files
in the main directory of .m files and in the subdirectory
lpoly. (Any file whose name begins "lpoly" is in the subdirectory
lpoly.)
auto17.m: This is the main driver program. You probably will not
call it directly, but all other ebbs functions call it.
autobic.m: EBBS for bivariate regression functions (under development)
ebbslogit.m: EBBS for local logistic regression
ebbsmvdef.m: Calls ebbsmvparam with default values of the parameters.
This is the function that most users will call. It
estimates the mean function or derivative thereof and
the variance function. It calls the following: autob17,
lpolydb, ebbsvar, lpolyres, bspan, ebbsmvparam, lpolyvar,
prebin
ebbsmvparam.m: Estimates the mean (or its derivatives)
ebbspam.m: Fits a semiparametric additive model, as discussed
in Opsomer and Ruppert, A root-n consistent backfitting
estimator for semiparametric additive modelling.
Preprint 97-20, Department of Statistics, Iowa State University
ebbsden.m: Calls ebbsdensity with default parameters
ebbsdensity.m: Estimates probability densities
ebbsvar.m: Estimates variance functions. Usually used indirectly
by calling ebbsmvdef or ebbsmvparam