R: compute deviance for cox model output

coxnet.deviance {glmnet}

R Documentation

compute deviance for cox model output

Given a fit or coefficients, compute the deciance (-2 log partial likelihood) for right-censored survival data

coxnet.deviance(pred = NULL, y, x = 0, offset = NULL,
  weights = NULL, beta = NULL)

`pred`	matrix of predictions
`y`	a survival response matrix, as produced by `Surv`
`x`	optional `x` matrix, if `pred` is `NULL`
`offset`	optional offset
`weights`	optional observation weights
`beta`	optional coefficient vector/matrix, supplied if `pred=NULL`

coxnet.deviance computes the deviance for a single prediction, or a matrix of predictions

a single or vector of deviances

Trevor Hastie
Maintainer: Trevor Hastie hastie@stanford.edu