
     Survival times of 358 ovarian cancer patients with information on
     three covariates,  karnofsky status (karn), tumor diameter (diam),
     figo stage (figo) and patients id.

datasets

A package for fitting Cox models with penalized ridge-type partial likelihood. The package includes functions for fitting simple Cox models with all covariates controlled by a ridge penalty. The weight of the penalty is optimised by using a REML type-algorithm. Models with time varying effects of the covariates can also be fitted. Some of the covariates may be allowed to be fixed and thus not controlled by the penalty. There are three different penalty functions, ridge, dynamic and weighted dynamic. Time varying effects can be fitted without the need of an expanded dataset.

Aris Perperoglou

CoxRidge

Cox Models with Dynamic Ridge Penalties

ova function

A data frame with 358 observations on the following 6 variables.
  <dl>
    <dt><code>time</code></dt><dd>Survival times in days.</dd> <dt><code>death</code></dt><dd>Status indicator, 0=censored, 1=death.</dd> <dt><code>karn</code></dt><dd>Karnofsky status at the start of the follow up.</dd> <dt><code>figo</code></dt><dd>Figo stage.</dd> <dt><code>diam</code></dt><dd>Timour diameter.</dd> <dt><code>x</code></dt><dd>Patient id.</dd> </dl> 

Format

Verweij, P. J. M. and Van Houwelingen H. C. (1993)
     <em>Cross-validation in survival analysis</em>. Statistics in Medicine:
     12, 2305-2314

ova: Ovarian cancer data set

Description

Usage

Arguments

Format

Source

Examples