[R] Estimating confidence intervals from several simulations

Zilefac Elvis zilefacelvis at yahoo.com
Wed May 8 17:01:45 CEST 2013


Hi,

I have 1000 simulations each arranged beneath another. The attached data is on monthly basis. First column is number of simulation, second is year and the rest, months.

Problem: 
1) Take sim 1 for example to be my original data from 1961-2005 with twelve months. You can delete the years. I want to obtain results with same dimensions as sim one but estimating the values in each row from the 1000 simulations as is the case with bootstrapping. Instead of bootstrapping, I have simulations from my model.
For example, after every 45th row, estimate values of row 1 and so on for other rows.

2) estimate confidence intervals for the resulting 45 rows by 12 columns matrix. Each value in each month and year should have a confidence interval.  OR you could estimate confidence intervals for each value in the data attached and then do task 1.

3) Put the confidence intervals in 45 by 12 matrix, one matrix for lower CIs and another 45 by 12 matrix for upper CIs.

Thanks so much for your great help.
Atem.


More information about the R-help mailing list