bst: Gradient Boosting

The package contains HingeBoost for binary and multi-class classification, with unequal misclassification costs for binary case. Functional gradient descent algorithm to optimize the hinge loss. The algorithm can fit linear and nonlinear classifiers.

Version: 0.3-2
Depends: rpart, gbm
Published: 2012-01-17
Author: Zhu Wang
Maintainer: Zhu Wang <zwang at ccmckids.org>
License: GPL-2
NeedsCompilation: no
In views: MachineLearning
CRAN checks: bst results

Downloads:

Package source: bst_0.3-2.tar.gz
MacOS X binary: bst_0.3-2.tgz
Windows binary: bst_0.3-2.zip
Reference manual: bst.pdf
News/ChangeLog:NEWS
Old sources: bst archive

Reverse dependencies:

Reverse suggests: caret, fscaret