Advertisement
Guest User

Untitled

a guest
Jul 21st, 2017
65
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 0.63 KB | None | 0 0
  1. 1.Repeat 1000 times.
  2.  
  3. 2. Generate $phi$ ( between .5 and .9) and sigma^2 (between 1.1 and 2).
  4. 3. Repeat 100 times:
  5. 4. Simulate an AR1 series with parameters $phi$ and
  6. $sigma^2$.
  7. 5. Estimate 95 % CI for these 2 parameters using dlmMLE ( from the dlm package in R)
  8. 6. Count the number of times that the confidence interval for each original parameter generated by the above method does NOT contain that parameter. This should be Binomial(size=100,p=.05).
  9. 7. Accumulate these counts in a vector. This should be 1000 realizations of Binomial(100,.05).
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement