ROCt: Time-Dependent ROC Curve Estimators and Expected Utility Functions

Contains functions in order to estimate diagnostic and prognostic capacities of continuous markers. More precisely, one function concerns the estimation of time-dependent ROC (ROCt) curve, as proposed by Heagerty et al. (2000) <doi:10.1111/j.0006-341X.2000.00337.x>. One function concerns the adaptation of the ROCt theory for studying the capacity of a marker to predict the excess of mortality of a specific population compared to the general population. This last part is based on additive relative survival models and the work of Pohar-Perme et al. (2012) <doi:10.1111/j.1541-0420.2011.01640.x>. We also propose two functions for cut-off estimation in medical decision making by maximizing time-dependent expected utility function. Finally, we propose confounder-adjusted estimators of ROC and ROCt curves by using the Inverse Probability Weighting (IPW) approach. For the confounder-adjusted ROC curve (without censoring), we also proposed the implementation of the estimator based on placement values proposed by Pepe and Cai (2004) <doi:10.1111/j.0006-341X.2004.00200.x>.

Package details

AuthorY. Foucher, E. Dantan, F. Le Borgne, and M. Lorent
MaintainerY. Foucher <Yohann.Foucher@univ-nantes.fr>
LicenseGPL (>= 2)
Version0.9.5
URL www.r-project.org www.divat.fr
Package repositoryView on CRAN
Installation Install the latest version of this package by entering the following in R:
install.packages("ROCt")

Try the ROCt package in your browser

Any scripts or data that you put into this service are public.

ROCt documentation built on May 2, 2019, 3:25 p.m.