Advertisement
Guest User

Untitled

a guest
Jun 27th, 2019
141
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 1.47 KB | None | 0 0
  1. fit = fitdistr(Y, "weibull")
  2. Warning message:
  3. In densfun(x, parm[1], parm[2], ...) : NaNs produced
  4. > warnings(fit)
  5. Warning message:
  6. In densfun(x, parm[1], parm[2], ...) : NaNs produced Error in
  7. at(list(...), file, sep, fill, labels, append) :
  8. argument 2 (type 'list') cannot be handled by 'cat'
  9.  
  10. shape scale
  11. 2.1103684 1537.2344072
  12. (0.2245888) (112.1596367)
  13.  
  14. Weibull Scale 1550.559
  15. Weibull Shape 2.1195
  16.  
  17. > library(Renext) ## for concentrated log-lik
  18. > try(fweibull(Y)) ## error (numerical pb with information matrix)
  19. > fit <- fweibull(Y / 1000) ## works
  20. > ## set parameters and logLik back to original scale
  21. > fit$est * c(1, 1000)
  22. shape scale
  23. 2.126225 1563.094460
  24.  
  25. > fit$sd * c(1, 1000)
  26. shape scale
  27. 0.2444308 114.1293266
  28.  
  29. > fit$loglik - length(Y) * log(1000)
  30. [1] -362.2237
  31.  
  32. > library(MASS)
  33. > ## set parameters and logLik back to original scale
  34. > fit2 <- fitdistr(Y / 1000, "weibull")
  35. > fit2$est * c(1, 1000)
  36. shape scale
  37. 2.126231 1563.095165
  38.  
  39. > fit2$sd * c(1, 1000)
  40. shape scale
  41. 0.2288605 114.9071653
  42.  
  43. > fit2$loglik - length(Y) * log(1000)
  44. [1] -362.2237
  45.  
  46. > survreg(Surv(Y)~1)
  47. Call:
  48. survreg(formula = Surv(Y) ~ 1)
  49.  
  50. Coefficients:
  51. (Intercept)
  52. 7.354423
  53.  
  54. Scale= 0.4703164
  55.  
  56. Loglik(model)= -362.2 Loglik(intercept only)= -362.2
  57. n= 46
  58. > exp(7.354423) # exponentiate the Intercept
  59. [1] 1563.095
  60. > 1/0.4703164 # take inverse of the Scale
  61. [1] 2.126228
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement