fic: Focused Information Criteria for Model Comparison

Compares how well different models estimate a quantity of interest (the "focus") so that different models may be preferred for different purposes. Comparisons within any class of models fitted by maximum likelihood are supported, with shortcuts for commonly-used classes such as generalised linear models and parametric survival models. The methods originate from Claeskens and Hjort (2003) <doi:10.1198/016214503000000819> and Claeskens and Hjort (2008, ISBN:9780521852258).

Version: 1.0.0
Depends: R (≥ 2.10)
Imports: stats, numDeriv, mvtnorm, ggplot2, scales, survival, tensor, abind
Suggests: knitr, rmarkdown, testthat, msm (≥ 1.6.6), flexsurv, sn, gapminder, GGally
Published: 2019-04-13
Author: Christopher Jackson [cre, aut] (package design and programming), Gerda Claeskens [aut] (method development and design advice), Howard Thom [ctb]
Maintainer: Christopher Jackson <chris.jackson at mrc-bsu.cam.ac.uk>
BugReports: https://github.com/chjackson/fic/issues
License: GPL-3
URL: https://github.com/chjackson/fic
NeedsCompilation: no
CRAN checks: fic results

Documentation:

Reference manual: fic.pdf
Vignettes: Different models for different purposes: focused model comparison in R
Examples of focused model comparison: linear regression
Focused model comparison with alternative loss functions
Examples of focused model comparison: multi-state models
Examples of focused model comparison: skew-normal models
Examples of focused model comparison: parametric survival modelling

Downloads:

Package source: fic_1.0.0.tar.gz
Windows binaries: r-devel: fic_1.0.0.zip, r-release: fic_1.0.0.zip, r-oldrel: fic_1.0.0.zip
macOS binaries: r-release (arm64): fic_1.0.0.tgz, r-oldrel (arm64): fic_1.0.0.tgz, r-release (x86_64): fic_1.0.0.tgz

Linking:

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