Plotting a imagesc matrix with 5 different colors and categories.

Hi all,

I want to compare the two tables attached (their values excluded NaNs) producing a plot that resembles that of imagesc with more colors namely for true positives (1 in both tables), true negatives (0 in both tables), false positives (1 in table rec and 0 in table original), false negatives (0 in table rec and 1 in table original) and NaNs (left black or grey maybe).

How can I do so?

The result should be a matrix with the dimensions of the ones provided with elements colored according to the group they belong to (see the above).

Thank you,

Federico Nutarelli

Answer 1

[Karthik]

Hello Giorgio

I understand you want to compare two tables and generate new matrices based on following conditions.

true positives - 1 in both tables

true negatives - 0 in both tables

false positives - 1 in table rec and 0 in table original

false negatives - 0 in table rec and 1 in table original

NaN - if NAN or blank on both sides

Following is the code that can compare two tables to create resultant matrix

% Convert tables to matrices assuming they are already imported

recMatrix = table2array(rec_table_normalized);

originalMatrix = table2array(original_mat2014);

% Check if matrices have the same dimensions

if size(recMatrix) ~= size(originalMatrix)

error('Tables must have the same dimensions.');

end

% Create logical matrices for different groups

truePositives = recMatrix == 1 & originalMatrix == 1;

trueNegatives = recMatrix == 0 & originalMatrix == 0;

falsePositives = recMatrix == 1 & originalMatrix == 0;

falseNegatives = recMatrix == 0 & originalMatrix == 1;

nanVals = isnan(recMatrix) | isnan(originalMatrix);

% Create a matrix to represent the groups

colorMatrix = zeros(size(recMatrix));

colorMatrix(truePositives) = 1; % Set true positives to 1 (or any other value/color)

colorMatrix(trueNegatives) = 2; % Set true negatives to 2 (or any other value/color)

colorMatrix(falsePositives) = 3; % Set false positives to 3 (or any other value/color)

colorMatrix(falseNegatives) = 4; % Set false negatives to 4 (or any other value/color)

colorMatrix(nanVals) = 0;

disp(colorMatrix)

% imagesc(colorMatrix);

Thanks,

Karthik.

[Giorgio Gnecco]

Thank you,

Federico Nutarelli (on behalf of Giorgio Gnecco)