Weibull-Shifted model


[weibullshifted_ex1]

   model
   {
      for( i in 1 : N )
      {
      x[i] ~ dweib3(nu, lambda, x0)      
      }
      
   # Prior distributions of the model parameters
   nu~ dunif(0, 5)
         lambda~ dunif(0, 5)   
         xMin <- ranked(x[], 1)x0 ~ dunif(0, xMin)      
   }
   
Data generated from Weibull shifted distribution with shape(nu)=0.75, scale(lambda) = 0.75 and location(x0) = 5.0

Data
list(N=50, x=c(6.137586, 7.074556, 9.894666, 5.013768, 7.428222, 5.833135,
5.061419, 8.666846, 5.888522, 8.503873, 6.434743, 5.264158, 5.019024, 5.959139,
7.992135, 5.883465, 5.168402, 8.979449, 5.701817, 7.072784, 8.699820, 5.369439, 10.756600, 5.670287, 5.483465, 5.653821, 5.389355, 5.045285, 5.359706, 7.800460, 6.922263, 13.273442, 5.046622, 23.082790, 6.073328, 5.185828, 5.721390, 6.625921, 5.963122, 8.283738, 5.219268, 5.427565, 7.062516, 8.093987, 5.143574, 8.968071, 7.455459, 5.060323, 5.329954, 5.203936))
Inits for chain 1
list(nu = 0.1, lambda = 0.1, x0 = 1.0)
   
Inits for chain 2
list(nu=1.0, lambda= 2.0, x0= 3.0)



Results

[weibullshifted_ex2]

[weibullshifted_ex3]

Compile 10 chains and do 100000 updates. Convergence bad but for chain with lowest deviance get lambda = 0.763452, nu =0.690463, x0 = 5.01, deviance = 156.509.

On this page