[R] timereg

Silvano silvano at uel.br
Mon Aug 5 20:27:36 CEST 2013


Hi,

I tried to fit a model using a timecox function in R version 
3.0.0, using sTRACE data  present in survival package.

If I use the 2.9.0 version, I don't have problems, but in 
3.0.0 version, I get the following error message:

> out <- timecox(Surv(time/365, status==9) ~ 
> age+sex+diabetes+chf+vf, sTRACE, max.time=7, n.sim=500)

Error in .C("OStimecox", as.double(times), 
as.integer(Ntimes), as.double(designX),  :
  "OStimecox" not available for .C() for package "timereg"

I don't know what's happend.

I would like use 3.0.0 version, but I don't know what's 
wrong. Somebody help me?


Also, when I use the aalen model I get differents results in 
3.0.0 and 2.9.0 versions. Why?

Version 3.0.0:

> fit1.semi <- aalen(Surv(time/365, 
> status==9)~age+sex+diabetes+chf+vf, 
> sTRACE,max.time=7,n.sim=500)
> summary(fit1.semi)

Additive Aalen Model

Test for nonparametric terms

Test for non-significant effects
            Supremum-test of significance p-value H_0: 
B(t)=0
(Intercept)                          7.29 
0.000
age                                  8.63 
0.000
sex                                  2.95 
0.060
diabetes                             2.31 
0.240
chf                                  5.30 
0.000
vf                                   2.95 
0.042

Test for time invariant effects
                  Kolmogorov-Smirnov test p-value 
H_0:constant effect
(Intercept)                       0.68600 
0.006
age                               0.00934 
0.008
sex                               0.16900 
0.078
diabetes                          0.22100 
0.184
chf                               0.14800 
0.176
vf                                0.46100 
0.008


in 2.9.0 version:

> fit1.semi <- aalen(Surv(time/365, 
> status==9)~age+sex+diabetes+chf+vf, 
> sTRACE,max.time=7,n.sim=500)
> summary(fit1.semi)
Additive Aalen Model

Test for nonparametric terms

Test for non-significant effects
            sup|  hat B(t)/SD(t) | p-value H_0: B(t)=0
(Intercept)                   7.29               0.000
age                           8.63               0.000
sex                           2.95               0.052
diabetes                      2.31               0.246
chf                           5.30               0.000
vf                            2.95               0.026

Test for time invariant effects
            sup| B(t) - (t/tau)B(tau)| p-value H_0: B(t)=b t
(Intercept)                    0.68600                 0.004
age                            0.00934                 0.004
sex                            0.16900                 0.084
diabetes                       0.22100                 0.208
chf                            0.14800                 0.158
vf                             0.46100                 0.004

part of them...


Thanks a lot,

--------------------------------------
Silvano Cesar da Costa
Departamento de Estatística
Universidade Estadual de Londrina
Fone: 3371-4346



More information about the R-help mailing list