LGDtoolkit: Collection of Tools for LGD Rating Model Development

The goal of this package is to cover the most common steps in Loss Given Default (LGD) rating model development. The main procedures available are those that refer to bivariate and multivariate analysis. In particular two statistical methods for multivariate analysis are currently implemented – OLS regression and fractional logistic regression. Both methods are also available within different blockwise model designs and both have customized stepwise algorithms. Descriptions of these customized designs are available in Siddiqi (2016) <doi:10.1002/9781119282396.ch10> and Anderson, R.A. (2021) <doi:10.1093/oso/9780192844194.001.0001>. Although they are explained for PD model, the same designs are applicable for LGD model with different underlying regression methods (OLS and fractional logistic regression). To cover other important steps for LGD model development, it is recommended to use 'LGDtoolkit' package along with 'PDtoolkit', and 'monobin' (or 'monobinShiny') packages. Additionally, 'LGDtoolkit' provides set of procedures handy for initial and periodical model validation.

Version: 0.2.0
Depends: R (≥ 2.10)
Imports: dplyr, monobin
Published: 2023-05-30
DOI: 10.32614/CRAN.package.LGDtoolkit
Author: Andrija Djurovic [aut, cre]
Maintainer: Andrija Djurovic <djandrija at gmail.com>
License: GPL (≥ 3)
URL: https://github.com/andrija-djurovic/LGDtoolkit
NeedsCompilation: no
Materials: README NEWS
CRAN checks: LGDtoolkit results

Documentation:

Reference manual: LGDtoolkit.pdf

Downloads:

Package source: LGDtoolkit_0.2.0.tar.gz
Windows binaries: r-devel: LGDtoolkit_0.2.0.zip, r-release: LGDtoolkit_0.2.0.zip, r-oldrel: LGDtoolkit_0.2.0.zip
macOS binaries: r-release (arm64): LGDtoolkit_0.2.0.tgz, r-oldrel (arm64): LGDtoolkit_0.2.0.tgz, r-release (x86_64): LGDtoolkit_0.2.0.tgz, r-oldrel (x86_64): LGDtoolkit_0.2.0.tgz
Old sources: LGDtoolkit archive

Linking:

Please use the canonical form https://CRAN.R-project.org/package=LGDtoolkit to link to this page.