Hi,
I need a help in a survival analysis using survreg function with weibull
distribution from survival package.
Look the data sample:
########## Start of script
dados <- structure(list(TFD = c(20L, 34L, 1L, 2L, 3L, 3L, 50L, 26L, 1L,
50L, 21L, 3L, 13L, 11L, 22L, 50L, 50L, 1L, 50L, 9L, 50L, 1L,
13L, 50L, 50L, 1L, 6L, 50L, 50L, 50L, 36L, 3L, 46L, 10L, 50L,
1L, 18L, 3L, 36L, 37L, 50L, 7L, 1L, 1L, 7L, 24L, 4L, 50L, 12L,
17L), Censor = c(1, 1, 1, 1, 1, 1, 0, 1, 1, 0, 1, 1, 1, 1, 1,
0, 0, 1, 0, 1, 0, 1, 1, 0, 0, 1, 1, 0, 0, 0, 1, 1, 1, 1, 0, 1,
1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 0, 1, 1)), .Names = c("TFD",
"Censor"), class = "data.frame", row.names = c(NA, -50L))
summary(dados)
TFD Censor
Min. : 1.00 Min. :0.00
1st Qu.: 3.00 1st Qu.:0.00
Median :17.50 Median :1.00
Mean :23.08 Mean :0.72
3rd Qu.:50.00 3rd Qu.:1.00
Max. :50.00 Max. :1.00
### TFD is the time for death
attach(dados)
m <- survreg(Surv(TFD,Censor)~1)
summary(m)
#Call:
#survreg(formula = Surv(TFD, Censor) ~ 1)
# Value Std. Error z p
#(Intercept) 3.466 0.250 13.87 9.67e-44
#Log(scale) 0.403 0.144 2.81 4.98e-03
#
#Scale= 1.5
#
#Weibull distribution
#Loglik(model)= -156.2 Loglik(intercept only)= -156.2
#Number of Newton-Raphson Iterations: 5
#n= 50
### Calculating the mean time for death, the time that 50% of population
is dead.
mu <- exp(3.466)
mu
#[1] 32.00845
### In this example the mean time for death (\mu) is 32.00845.
### Now the alpha calculation based on scale parameter
alpha <- 1/1.5
alpha
#[1] 0.6666667
### Making the curve
curve(exp((-mu^(-alpha))*(x^alpha)),from=0,to=50,ylim=c(0,1),col=1,ylab="Survival",xlab="Time")
### Adding lines in graphic to show the point corresponding to the
coordinates, y=0.5 (50% of mortality) and x=32 that is the time to death
for 50% of population.
abline(h=0.5,lty=2)
abline(v=mu,lty=2)
points(mu,0.5,cex=2)
The problem is, I expected that the estimated curve from weibull cross
this point(32,0.5) but it is not really.
What is the problem with this analysis?
Thanks
Ronaldo
--
1ª lei - Suas férias começam após a defesa e entrega de sua dissertação.
--Herman, I. P. 2007. Following the law. NATURE, Vol 445, p. 228.
> Prof. Ronaldo Reis Júnior
| .''`. UNIMONTES/DBG/Lab. Ecologia Comportamental e Computacional
| : :' : Campus Universitário Prof. Darcy Ribeiro, Vila Mauricéia
| `. `'` CP: 126, CEP: 39401-089, Montes Claros - MG - Brasil
| `- Fone: (38) 3229-8192 | ronaldo.reis@unimontes.br
| http://www.ppgcb.unimontes.br/lecc | LinuxUser#: 205366
[[alternative HTML version deleted]]
