pga: An ensemble method for variable selection by carrying out Darwinian evolution in parallel universes

PGA is an ensemble algorithm similar in spirit to AdaBoost and random forest. It can "boost up" the performance of "bad" selection criteria such as AIC and GCV. (Technometrics 48, 491-502; The American Statistician 62, 97-109, Sec 4.3)

Version: 0.1-1
Date: 2007-8-15
Author: Dandi Qiao, Mu Zhu
Maintainer: Mu Zhu <m3zhu at math.uwaterloo.ca>
License: GPL (≥2)
In views: ChemPhys
CRAN checks: pga results

Downloads:

Package source: pga_0.1-1.tar.gz
MacOS X binary: pga_0.1-1.tgz
Windows binary: pga_0.1-1.zip
Reference manual: pga.pdf