Trouble calculating column mean from a matrix - using different indices for each row

Hello,

I am trying to calculate a column of means from a 2-D matrix. The length of the mean is equally to the number of rows of the matrix.

However, the index to each row is different. For example, row 1 I want the mean of colums 4 to 8. Row 2 I want the means of columns 3 to 5.

I have tried

mean(myMatrix(:,VariableIndex1:VaribleIndex2))

Where VariableIndex1 and VariableIndex2 are columns for the start and end indices of the columns I need means from

For the example above this would be:

mean(myMatrix(:,[4;3]:[3:5]))

- but this seems to give a non-sensical result.

Any thoughts or advice on how to do this? The column acutally has 100 000 rows so I don't want to loop through it.

Thanks.

Answer 1

[Matt J]

[m,n]=size(myMatrix); %Edited
e=1:n;

mask=VariableIndex1(:)<=e & e<=VariableIndex2(:);

result=sum(myMatrix.*mask,2)./sum(mask,2);

[David McVea]

Hi - perfect, thanks. Except note the first line should be

[m,n]=size(myMatrix);

[Matt J]

Right you are!