smoothSurvReg.fit {smoothSurv}

Work Function to Fit the Model Using 'smoothSurvReg'

Description

Fit the survival regression model with smoothed error distribution. This function is not to be called by the user.

Usage

smoothSurvReg.fit(x, z, y, offset = NULL, correctlik, init, controlvals, common.logscale)

Arguments

x

A covariate matrix.

z

A covariate matrix for log(scale).

y

A two- or three- column matrix with response. The last column indicate the type of censoring. See Surv for details.

offset

A vector with possible offset term.

correctlik

Correction term to the likelihood due to the log transformation of the response.

init

A list with components beta, scale, ccoef giving the initial values for the fitting process.

controlvals

A list returned by smoothSurvReg.control.

common.logscale

Indicator (TRUE/FALSE) indicating whether the log-scale is same for all observations or whether it depends on covariates.

Value

A list with components regres, spline, loglik, aic, degree.smooth, var, var2, dCdC, iter, estimated, warning, fail, H, I, G, U.

Note

WARNING: Most users will call the higher level routine smoothSurvReg. Consequently, this function has very few error checks on its input arguments.

Author(s)

Arnošt Komárek [e-mail]