how to select higher value in a matrix and change that by introducing an error?

I have a matrix,

a= [ 0.83 0.85 0.97 0.1 0.95 0.93 0.2 ; 0.2 0.12 0.12 0.76 0.77 0.78 0.25 ; 0.88 0.32 0.11 0.77 0.87 0.89 0.99]

by ceil(a) we get c matrix: c=[ 1 1 1 0 1 1 0 ; 0 0 0 1 1 1 0 ; 1 0 0 1 1 1 1]

now i want to introduce error position wise as err=1:3

so, at first err=1 will be introduced to first row where the only highest value is find in matrix a and 1 will become 0,same wise in 2nd row and 3rd row. Then err=2, that means the first two highest value is find and 1 will become zero, then same will be carried out for 2nd,3rd row. Next err=3...where 3 higher value will be changed.

please help me out asap....

plz plz plz.....

[J. van Delft]

First of all, the ceil command rounds the elements of a to the nearest integers greater than or equal to a. So your matrix c will consist of only 1's. Instead you could use "round" to produce the results you want.

Next, the "max" command is able to give you the value and position of the maximum value of your matrix. For example, for the 1st row of a:

[val,pos] = max(a(1,:));

This will give you val = 0.97, pos = 3. Then simply change the value of the element corresponding to pos = 3 in the c-matrix:

c(1,pos) = 0

A for-loop can reproduce the same results for the 2nd and 3rd row.

After this you can repeat the process to find the second highest value. Don't forget to change the highest value you found before to zero or else you'll find the same result over and over again.

[suchismita]

ok....thank you....

can't i find 2 highest value together??????

[the cyclist]

You can use the sort() command rather than the max() command to find multiple values.

[suchismita]

in sort() the position of matrix is changing...i want to identify the positions so that i can convert 1 to 0

[Roger Wohlwend]

Yes, but the sort command also tells you the original position of the sorted values (it's the second output variable). Use that information.

Answer 1

[George Papazafeiropoulos]

Are you looking for something like this?

a= [ 0.83 0.85 0.97 0.1 0.95 0.93 0.2;
    0.2 0.12 0.12 0.76 0.77 0.78 0.25;
    0.88 0.32 0.11 0.77 0.87 0.89 0.99];
c=round(a)
for i=1:3
    for err=1:3
        [~,d]=sort(a(err,:),'descend');
        a(err,d(1:i))=0;
    end
end
cfinal=round(a)

I hope this will help!

[suchismita]

thank u so much....how can i save err 1,2,3 example separately...?????

[suchismita]

sort is changing the 'a' matrix itself....but i dont want to disturb a matrix, the changes i want to make in c....as in matrix a in first row 0.97 is highest so that value in c will become zero same wise in 2nd row 0.77 (a(2,5)) will become zero, same process will be carried in 3rd row. In second round, first row will take the two highest numbers as 0.97 (1,3)and next highest to it is 0.95 (1,5) will become zero.... this is the concept....plz help me....

[George Papazafeiropoulos]

I think the following applies if you want the a matrix unaltered. The outer for loop (for i=1:3) is not required in this case, because for all i (=1,2,3) the maximum element of each row of A is selected and set to zero, so we can set i=3, select the three highest values, set them to zero and get rid of the initial i loop:

a= [ 0.83 0.85 0.97 0.1 0.95 0.93 0.2;
    0.2 0.12 0.12 0.76 0.77 0.78 0.25;
    0.88 0.32 0.11 0.77 0.87 0.89 0.99];
c=round(a)
for err=1:3
    [~,d]=sort(a(err,:),'descend');
    c(err,d(1:3))=0;
end

Answer 2

[suchismita]

thank u so much....ya i got it now....thank u so so so much...