How can i find the RMSE between 2D double matrices?

Can you please help me and tell me how to find the RMSE between two matrices. I have one matrice of satellite data with dimensions 2521x2889 double, and a matrice of model data with exact the same dimensions, 2521x2889 double. I also have the latitude and the longitude matrices. I want to find the RMSE between the satellite data and the model data by making a new, exact same dimensioned matrice with the RMSE values in order to plot it along with the longitude and the latitude. I have tried {RMSerror_f1_20150201=sqrt( sum(sum((model_data - sat_data).^2 ))/(2521*3889))} but it gives me a new NaN matrice. Thank you in advance!

Answer 1

[Greg Heath]

 E    = T-Y;
 SQE  = E.^2
 MSE  = mean(SQE(:))
 RMSE = sqrt(MSE)

Hope this helps.

Thank you for formally accepting my answer

Greg

[Panourgias Siderakos]

Thank you very much for the help!!!

[KSSV]

@ Panourgias Siderakos can you explain the difference between both the answers?

Answer 2

[KSSV]

Let A and B be your matrices.

iwant = sqrt(mean(A(:).^2 - B(:).^2)) ;

[Panourgias Siderakos]

It gives me a matrice with a 1x1 double dimensions. Also, in the workspace, this matrice has value of NaN.I don't know why but i am sure that's not the desired outcome.

[KSSV]

Refer more here: https://in.mathworks.com/matlabcentral/answers/4064-rmse-root-mean-square-error

Any data in your matrix is NaN?

[Panourgias Siderakos]

Yes.The values where land is (only in the satellite data) are NaN. All the others are values of sea surface temperature in degrees Celsius.

[KSSV]

Ohh...then obviously you will get NaN...you have to remove these NaN's then..

Let A be your satellite data which have NaN's. Follow like below:

A = A(~isnan(A)) ;
B = B(~isnan(A)) ;
iwant = sqrt(mean(A.^2 - B.^2)) ;

[Panourgias Siderakos]

Oh God.Sorry for asking but i am still learning a lot here.How can i do that?