VRPM: Visualizing Risk Prediction Models

This is a package to visualize risk prediction models. For each predictor, a color bar represents the contribution to the linear predictor or latent variable. A conversion from the linear predictor to the estimated risk or survival is also given. (Cumulative) contribution charts enable to visualize how the estimated risk for one particular observation is obtained by the model. Several options allow to choose different color maps, and to select the zero level of the contributions. The package is able to deal with 'glm', 'coxph', 'mfp', 'multinom' and 'ksvm' objects. For 'ksvm' objects, the visualization is not always exact. Functions providing tools to indicate the accuracy of the approximation are provided in addition to the visualization.

Version: 1.1
Imports: Hmisc, survival, fields, R2HTML, viridis, kernlab, ROCR, ggplot2, shiny, methods, stats, grDevices, graphics, utils
Suggests: mfp, VGAMdata, nnet
Published: 2016-09-01
Author: Vanya Van Belle
Maintainer: Vanya Van Belle <vanya.vanbelle at esat.kuleuven.be>
License: GPL-3
NeedsCompilation: no
Materials: NEWS
CRAN checks: VRPM results

Downloads:

Reference manual: VRPM.pdf
Package source: VRPM_1.1.tar.gz
Windows binaries: r-devel: VRPM_1.1.zip, r-release: VRPM_1.1.zip, r-oldrel: VRPM_1.1.zip
OS X Mavericks binaries: r-release: VRPM_1.1.tgz, r-oldrel: VRPM_1.1.tgz

Linking:

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