Question

I’m creating a function to use multiple matrices in an analysis studio.

The matrices are given with the same name with a date reference in the name (month to month and year to year: nov-1956 is matrix5611, dec-1956 is matrix5612, jan-1957 is matrix5712, and so on until the end of 1999.

For each one there should be a comparison between the mean value of each month/year (depending of what area of study are you focused on).

I’m trying to use some loops to vary the name of the input matrix instead of write manually date by date, but a function that helped would be useful.

Any idea or useful function?

Answer 1

If you have your data in different matrices, you can use eval to store the means to some matrix, in this example MeanMatrix, in which Y dimension is year and X dimension is month:

Edit: It’s not running number from 5611 but yymm…

Edit: It seems that matrices don’t begin from January 1956 but from November 1956.

% add here missing months matrix index strings.

MissingMatricesCellArray = {'5601', '5602', '5603', '5604', '5605', '5606', '5607', '5608', '5609', '5610'};

% MissingmatricesCellArray = {};

for Year = 56:99
    for Month = 1:12
        NumString = sprintf('%02d%02d', Year, Month);

        % calculate and store means only for matrices that are not missing.
        if ~(ismember (cellstr(NumString), MissingMatricesCellArray))
            MeanMatrix(Year,Month) = mean(mean(eval ([ 'matrix', NumString ])));
        end
     end
end

Then you can compare the means of months and years the way you wish.