Predicted Values for a 'survreg' Object

Description:

     Predicted values for a 'survreg' object

Usage:

     ## S3 method for class 'survreg':
     predict(object, newdata,  
     type=c("response", "link", "lp", "linear", "terms", "quantile",  
             "uquantile"),
      se.fit=FALSE, terms=NULL, p=c(0.1, 0.9), na.action=na.pass, ...)

Arguments:

  object: result of a model fit using the 'survreg' function.  

 newdata: data for prediction.  If absent predictions are for the 
          subjects used in the original fit.  

    type: the type of predicted value.   This can be on the original
          scale of the data (response),  the linear predictor
          ('"linear"', with '"lp"' as an allowed abbreviation),  a
          predicted quantile on the original scale of the data
          ('"quantile"'),  a quantile on the linear predictor scale
          ('"uquantile"'), or the matrix of terms for the linear
          predictor ('"terms"'). At this time '"link"' and linear
          predictor ('"lp"') are identical.  

  se.fit: if 'TRUE', include the standard errors of the prediction in
          the result.  

   terms: subset of terms.  The default for residual type '"terms"' is
          a matrix with  one column for every term (excluding the
          intercept) in the model.  

       p: vector of percentiles.  This is used only for quantile
          predictions.  

na.action: applies only when the 'newdata' argument is present, and
          defines the missing value action for the new data.  The
          default is to include all observations.

     ...: for future methods
Value:

     a vector or matrix of predicted values.

References:

     Escobar and Meeker (1992). Assessing influence in regression
     analysis with  censored data. _Biometrics,_ 48, 507-528.

Examples:

     # Draw figure 1 from Escobar and Meeker
     fit <- survreg(Surv(time,status) ~ age + age^2, data=stanford2, 
             dist='lognormal') 
     plot(stanford2$age, stanford2$time, xlab='Age', ylab='Days', 
             xlim=c(0,65), ylim=c(.01, 10^6), log='y') 
     pred <- predict(fit, newdata=list(age=1:65), type='quantile', 
                      p=c(.1, .5, .9)) 
     matlines(1:65, pred, lty=c(2,1,2), col=1) 

     # Predicted Weibull survival curve for a lung cancer subject with
     #  ECOG score of 2
     lfit <- survreg(Surv(time, status) ~ ph.ecog, data=lung)
     pct <- 1:98/100   # The 100th percentile of predicted survival is at +infinity
     ptime <- predict(lfit, newdata=data.frame(ph.ecog=2), type='quantile',
                      p=pct, se=TRUE)
     matplot(cbind(ptime$fit, ptime$fit + 2*ptime$se.fit,
                              ptime$fit - 2*ptime$se.fit)/30.5, 1-pct,
             xlab="Months", ylab="Survival", type='l', lty=c(1,2,2), col=1)

########################    Residuals. 

Compute Residuals for 'survreg' Objects

Description:

     This is a method for the function 'residuals' for objects
     inheriting from class 'survreg'.

Usage:

     ## S3 method for class 'survreg':
     residuals(object, type=c("response", "deviance","dfbeta","dfbetas",
     "working","ldcase","ldresp","ldshape", "matrix"), rsigma=TRUE,
     collapse=FALSE, weighted=FALSE, ...)

Arguments:

  object: an object inheriting from class 'survreg'. 

    type: type of residuals, with choices of '"response"',
          '"deviance"', '"dfbeta"', '"dfbetas"', '"working"',
          '"ldcase"', '"lsresp"', '"ldshape"', and '"matrix"'.  See the
          LaTeX documentation ('survival/doc/survival.ps.gz') for more
          detail. 

  rsigma: include the scale parameters in the variance matrix, when
          doing computations. (I can think of no good reason not to). 

collapse: optional vector of subject groups.  If given, this must be of
          the same length as the residuals, and causes the result to be
          per group residuals. 

weighted: give weighted residuals?  Normally residuals are unweighted. 

     ...: other unused arguments

Value:

     A vector or matrix of residuals is returned. Response residuals
     are on the scale of the original data, working residuals are on
     the scale of the linear predictor, and deviance residuals are on
     log-likelihood scale. The dfbeta residuals are a matrix, where the
     ith row gives the approximate change in the coefficients due to
     the addition of subject i. The dfbetas matrix contains the dfbeta
     residuals, with each column scaled by the standard deviation of
     that coefficient.

     The matrix type produces a matrix based on derivatives of the
     log-likelihood function.  Let L be the log-likelihood, p be the
     linear predictor X %*% coef, and s be log(sigma).  Then the 6
     columns of the matrix are L, dL/dp,ddL/(dp dp), dL/ds, ddL/(ds ds)
      and ddL/(dp ds).  Diagnostics based on these quantities are
     discussed in an article by Escobar and Meeker. The main ones are
     the likelihood displacement residuals for perturbation of a case
     weight ('ldcase'), the response value ('ldresp'), and the 'shape'.

References:

     Escobar, L. A. and Meeker, W. Q. (1992). Assessing influence in
     regression analysis with censored data. Biometrics:48,507-528.

Examples:

     fit <- survreg(Surv(time,status) ~x, aml)
     rr  <- residuals(fit, type='matrix')