README.md

Introduction of nomovalidation

nomovalidation is an R package to measure the prediction performance of a Cox PH model by internal calibration, internal validation, and external validation. This package was built on the R packages rms and survAUC. "nomo" in the package name comes from nomogram, which is a visualization of a Cox PH model. Physicians and clinicians use nomogram to quickly and roughly predict the survival of a patient based on the covariables (age, race, etc). Before using this package, a Cox PH model need to be developed using the training dataset. This package contains 3 R functions:

#### 1. get_internal_calibrarion.R: To investigate the predictive performance of the Cox PH model at a specific value of time-to-event you are interested in, a calibration plot will be generated. #### 2. get_internal_validation.R: To assess the performance of the developed Cox PH model by bootstraping the training dataset, Harrell's C-index will be calculated for internal validation. #### 3. get_external_validation.R: To further assess the performance of the developed Cox PH model by applying it to an independent testing dataset, Uno's C-index will be calculated for external validation.

Installation in R

1. Install R packages rms and survAUC

install.packages("rms")
install.packages("survAUC")

2. Install R package nomovalidation

install.packages("devtools")
library(devtools)
install_github("ccwangxi/nomovalidation")


ccwangxi/nomovalidation documentation built on Feb. 14, 2020, 9:21 a.m.