how to calculate and plot rmse matrix

Hi,im a beginner in matlab. I need to plot for a rmse matrix. i need to calculate the rmse and reshape it into a heatmap (a 8x5 matrix). How do I calculate the rmse of 40 models and one observation? i have listed out my models in the code, and am retrieving it from 'odat' for observation and 'gdat' for the 40 models. i've looked through all the rmse commands but i couldn't really understand, so need some help on it.

4 Comments

Don't worry about the commands now.
Try describing the mathematical formula of your rmse matrix first.
We'll see which one suits your application.
i have the data of the annual cycle of sea surface temperature produced by 40 climate models and will need to compare it with the data of an observation. i've done calculating and plotting the same thing for correlation matrix using:
[R,p]=corrcoef(odat,gdat);
Rval(g,1) = R(1,2);
pval(g,1) = p(1,2);
and reshape it into:
Rval_matrix = reshape(Rval,8,5);
pval_matrix = reshape(pval,8,5);
however now i would need to calculate the rmse and not sure how to do it
Have you explored rmse, introduced in MATLAB R2022b?
yes. i've tried using the E = rmse(F,A), but it only gives one value of rmse. How do i make sure i get values for all the models?

Sign in to comment.

 Accepted Answer

Let O be your observation data array of size mX1 and M be your model data array of size mx40.
O = rand(30,1) ;
M = rand(30,40) ;
RMSE = sqrt(mean((O - M).^2));
RMSE = reshape(RMSE,8,5) ;
pcolor(RMSE)
colorbar
You arrange all your models into an array and get them to the same dimensions of Observations.

3 Comments

thank you, i managed to plot like so. but, would you care to explain how do i determine the dimension size (m), as what you did is 30?
oh sorry, i realised that there's only a total of 28 boxes/models (7x4) when it should be 8x5. an example of what im expecting;
Just to show an example I have taken 30....it depends on your data. Hope you problem is solved?

Sign in to comment.

More Answers (0)

Community Treasure Hunt

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

Start Hunting!