95% uncertainty bounds

13 views (last 30 days)
B on 21 Mar 2022
Commented: B on 21 Mar 2022
Please can some one show me how to compute the 95% Prediction uncertainty (PPU) in MATLAB?
I have tried to look into fit, ci, predint but getting very confused about the fittype, and the results. I am not sure I am using the right functions or commands.
To be clear, I have 100 simulations per day for 44195 days and needs to estimate the interval around daily median or mean. My outputs are just time and paramenter.

Answers (1)

Dave B
Dave B on 21 Mar 2022
Edited: Dave B on 21 Mar 2022
If you have 100 simulations per day, and you want to use your simulations to estimate 2.5% to 97.5% around the median (a 95% confidence interval), you could rephrase that as the 2.5th percentile and 97.5th percentile of your simulations.
a = randn(100,44195);
botmedtop = prctile(a,[2.5 50 97.5]); % Lower CI, median, upper CI
plot(botmedtop') % note the transpose, plot will make one line per column and botmedtop is 3 x 44195
  1 Comment
B on 21 Mar 2022
Hi Dave B, thanks for the help. What I want is prediction uncertainty (interval) and not confidence interval (CI). But I perfectly agree with your approach.

Sign in to comment.




Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!