dynamichazard: Dynamic Hazard Models using State Space Models

Contains functions that lets you fit dynamic hazard models using state space models. The first implemented model is described in Fahrmeir (1992) <doi:10.1080/01621459.1992.10475232> and Fahrmeir (1994) <doi:10.1093/biomet/81.2.317>. Extensions hereof are available where the Extended Kalman filter is replaced by an unscented Kalman filter and other options including particle filters.

Version: 0.5.1
Depends: R (≥ 3.3.2), stats, graphics, utils, survival
Imports: parallel, Rcpp (≥ 0.12.6), boot, stringr, data.table
LinkingTo: Rcpp, RcppArmadillo
Suggests: testthat, knitr, rmarkdown, timereg, captioner, biglm, httr, mgcv, shiny, formatR, R.rsp, speedglm, dichromat, colorspace
Published: 2017-11-25
Author: Benjamin Christoffersen [cre, aut], Alan Miller [cph], Anthony Williams [cph], Boost developers [cph], R-core [cph]
Maintainer: Benjamin Christoffersen <boennecd at gmail.com>
BugReports: https://github.com/boennecd/dynamichazard/issues
License: GPL-2
URL: https://github.com/boennecd/dynamichazard
NeedsCompilation: yes
SystemRequirements: C++11
Materials: NEWS
In views: Survival
CRAN checks: dynamichazard results


Reference manual: dynamichazard.pdf
Vignettes: Comparing methods for time varying logistic models
Bootstrap illustration
Particle filters in the dynamichazard package
Package source: dynamichazard_0.5.1.tar.gz
Windows binaries: r-devel: dynamichazard_0.5.1.zip, r-release: dynamichazard_0.5.1.zip, r-oldrel: dynamichazard_0.5.1.zip
OS X binaries: r-release: dynamichazard_0.5.1.tgz, r-oldrel: dynamichazard_0.5.1.tgz
Old sources: dynamichazard archive


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