crq {crq} | R Documentation |
Fits a conditional quantile regression model for censored data.
crq(formula, data, subset, weights, na.action, method = "grid", contrasts = NULL, ...)
formula |
A formula object, with the response on the left of the `~'
operator, and the terms on the right. The response must be a
survival object as returned by the Surv function. |
data |
A data.frame in which to interpret the variables named in the `formula', in the `subset', and the `weights' argument. |
subset |
an optional vector specifying a subset of observations to be used in the fitting process. |
weights |
vector of observation weights; if supplied, the algorithm fits to minimize the sum of the weights multiplied into the absolute residuals. The length of weights vector must be the same as the number of observations. The weights must be nonnegative and it is strongly recommended that they be strictly positive, since zero weights are ambiguous. |
na.action |
a function to filter missing data. This is applied to the model.frame after any subset argument has been used. The default (with 'na.fail') is to create an error if any missing values are found. A possible alternative is 'na.omit', which deletes observations that contain one or more missing values. |
method |
The algorithmic method used for fitting. There are currently
two options. The default method is "grid" which fits the crq process
on a specified grid of tau values in (0,1), with default spacing
=min (0.01,1/(2n^{.7})), where n is the sample size.
The other method is "pivot" which fits the process on a grid
determined by the data; this method is analogous to fitting the
whole rq process with rq . |
contrasts |
a list giving contrasts for some or all of the factors default = 'NULL' appearing in the model formula. The elements of the list should have the same name as the variable and should be either a contrast matrix (specifically, any full-rank matrix with as many rows as there are levels in the factor), or else a function to compute such a matrix given the number of levels. |
... |
additional arguments for the fitting routine (see
crq.fit . |
The algorithm uses recursively reweighted quantile regression to estimate
linear conditional quantile functions for censored data as in Portnoy (2003).
If method is "pivot", all break points are computed. If method is "grid",
then estimates are computed only on a fixed grid, see crq.fit
for
details on how to specify this grid. Control parameters needed by
crq.fit
can be passed in crq
via the ...
argument.
For smaller data sets (n<400 or 500), using "pivot" is generally better
and sufficiently quick. For n>400, it is generally adequate to
use "grid" and let step be somewhat less that 1/sqrt(n).
see crq.object
Stephen Portnoy & Tereza Neocleous
Portnoy, Stephen (2003) Censored Regression Quantiles, JASA, 98, 1001-1010.
data(uis) #estimate the crq model using log(TIME) AFT specification fit <- crq(Surv(log(TIME), CENSOR) ~ ND1 + ND2 + IV3 + TREAT + FRAC + RACE + AGE * SITE, data = uis) Sfit <- summary(fit,1:19/20) PHit <- coxph(Surv(TIME, CENSOR) ~ ND1 + ND2 + IV3 + TREAT + FRAC + RACE + AGE * SITE, data = uis) plot(Sfit, CoxPHit = PHit)