| proprate2 {surv2sample} | R Documentation |
proprate2 estimates the two-sample proportional rate transformation model
(proportional hazards, proportional odds) for censored data using the simplified
partial likelihood.
proprate2(x, group, model = 0, beta.init = 0, maxiter = 20,
eps = 1e-09)
x |
a "Surv" object, as returned by the Surv
function. |
group |
a vector indicating to which group each observation belongs. May contain values 1 and 2 only. |
model |
the type of model. Possible values are 0 for proportional hazards, 1 for proportional odds. |
beta.init |
the initial parameter value for iteration. |
maxiter |
the maximum number of iterations. |
eps |
the convergence tolerance parameter. The convergence criterion
is |(l-l_old)/l|<eps. |
This function fits the proportional rate model for two samples of censored survival data. Currently two most important models are implemented: proportional hazards and proportional odds. The estimation procedure is based on a two-sample simplification of the partial for the two-sample situation, see Bagdonavicius and Nikulin (2000). (For proportional hazards, this method is the usual partial likelihood.)
A list of class "proprate2.fit" with main components:
beta |
the estimate. |
var |
its variance. |
G0 |
the cumulative baseline rate (at times time). |
time |
sorted times. |
iter |
the number of iterations used. |
converged |
logical. Did the iterations (appear to) converge? |
loglik.init |
the simplified partial likelihood at the initial value of the parameter. |
loglik |
the simplified partial likelihood at the estimate. |
d11 |
the derivative of the score. |
sigma11 |
variance of the score (for proportional hazards sigma11
equals d11). |
G1, G2 |
cumulative transformation rates computed separately
in the two groups (both of length n, at times time). |
David Kraus (http://www.davidkraus.net/)
Bagdonavicius, V. and Nikulin, M. (2000) On goodness-of-fit for the linear transformation and frailty models. Statist. Probab. Lett. 47, 177–188.
Kraus, D. (2007) Checking proportional rates in the two-sample transformation model. Research Report 2203, Institute of Information Theory and Automation, Prague. Available at http://www.davidkraus.net/surv2sample/.
There is a plot method for objects
returned by proprate2.
See proprate2.neyman, proprate2.ks,
proprate2.gs for tests of the proportional rate assumption.
## chronic active hepatitis data
data(hepatitis)
## fit the proportional odds model
fit = with(hepatitis, proprate2(Surv(time, status), treatment,
model = 1))
fit
## plot model-based and model-free estimates of odds functions
plot(fit)