predict.grpsurv {grpreg}

Model predictions for grpsurv objects

Description

Similar to other predict methods, this function returns predictions from a fitted grpsurv object.

Usage

## S3 method for class 'grpsurv'
predict(
  object,
  X,
  type = c("link", "response", "survival", "hazard", "median", "norm", "coefficients",
    "vars", "nvars", "groups", "ngroups"),
  lambda,
  which = 1:length(object$lambda),
  ...
)

Arguments

object

Fitted grpsurv model object.

X

Matrix of values at which predictions are to be made. Not required for some type values.

type

Type of prediction:

• link: linear predictors

• response: risk (i.e., exp(link))

• survival: the estimated survival function

• hazard: the estimated cumulative hazard function

• median: median survival time

• The other options are all identical to their grpreg() counterparts

lambda

Regularization parameter at which predictions are requested. For values of lambda not in the sequence of fitted models, linear interpolation is used.

which

Indices of the penalty parameter lambda at which predictions are required. Default: all indices. If lambda is specified, this will override which.

...

Not used.

Details

Estimation of baseline survival function conditional on the estimated values of beta is carried out according to the method described in Chapter 4.3 of Kalbfleisch and Prentice.

Value

The object returned depends on type.

Author(s)

Patrick Breheny

References

• Kalbfleish JD and Prentice RL (2002). The Statistical Analysis of Failure Time Data, 2nd edition. Wiley.

See Also

grpsurv()

Examples

data(Lung)
X <- Lung$X

y <- Lung$y
group <- Lung$group
 
fit <- grpsurv(X, y, group)
coef(fit, lambda=0.05)
head(predict(fit, X, type="link", lambda=0.05))
head(predict(fit, X, type="response", lambda=0.05))
 
# Survival function
S <- predict(fit, X[1,], type="survival", lambda=0.05)
S(100)
S <- predict(fit, X, type="survival", lambda=0.05)
plot(S, xlim=c(0,200))
 
# Medians
predict(fit, X[1,], type="median", lambda=0.05)
M <- predict(fit, X, type="median")
M[1:10, 1:10]
 
# Nonzero coefficients
predict(fit, type="vars", lambda=c(0.1, 0.01))
predict(fit, type="nvars", lambda=c(0.1, 0.01))