stratasphere: Stratified Spherical Approximate Conditional Modeling

For data where the observations occur or are collected in a stratified fashion, the strata level effect is sometimes not of interest or is simply impossible to estimate. Conditioning on the strata eliminates these difficulties, but leads to a combinatorially complex calculation. The spherical approximate method replaces this complex calculation with a closed form solution, leading to a simple fixed point iteration method for fitting the linear model. This package implements linear model fitting via Stratasphere(), and stepwise regression and boosting extensions via StratasphereStep() and StratasphereBoost(), respectively. Note that this package is distributed under the Apache 2.0 license.

Version: 1.0
Depends: survival, mboost, hash
Published: 2011-07-23
Author: Daniel Percival
Maintainer: Daniel Percival <dperciva at andrew.cmu.edu>
License: Apache License 2.0
CRAN checks: stratasphere results

Downloads:

Package source: stratasphere_1.0.tar.gz
MacOS X binary: stratasphere_1.0.tgz
Windows binary: stratasphere_1.0.zip
Reference manual: stratasphere.pdf