Question regards making a function in matlab involve matrices

Hi, I'm a completely newbies to Matlab, just got a student version a few days ago. And I was asked to solve this question using matlab. Can anyone guide me in making a function to satisfy the condition below.

Write a MATLAB function called noisify which accepts a 7 x 4 matrix of data and which returns a 7 x 4 matrix of corrupted data, where each column has only one bit randomly flipped to simulate an error in transmission occurring.
Hint: The MATLAB functions rand, round or randi may be useful.

Best Regards

Edit: I forgot to add more information about the question, the initial input is

[1 0 01; 0 0 1 0; 1 1 0 1; 0 0 1 1; 1 1 0 0; 0 1 1 1; 1 0 0 0]

a 7 by 4 matrice, my goal is to find a function that will corrupt this matrix while at the same time maintaining a 7x4 function and a binary number as output (0 or 1).

Is it possible for me to write a script like

function [C] = noisify([1 0 0 1; 0 0 1 0; 1 1 0 1; 0 0 1 1; 1 1 0 0; 0 1 1 1; 1 0 0 0])
%C is a 7 by 4 matrix which has been corrupted 

C= [1 0 0 1; 0 0 1 0; 1 1 0 1; 0 0 1 1; 1 1 0 0; 0 1 1 1; 1 0 0 0] - round(rand(7,4))
end

However I don't know how to limit the answer to number 0 or 1 only. How do I do that?

[Walter Roberson]

Functions need to take parameter names such as

function C = noisify( original_matrix )
  C = original_matrix - round(rand(7,4));
 end

[Samious]

I try that method but I still get some of the answer in the form of -1. Is there a way to script the function that the answer will only be 0s or 1s?

[Walter Roberson]

Consider the possibility that your expression is wrong :-)

If you are starting with a 0 and it flips, then how much do you need to subtract? If you are starting with a 1 and it flips, then how much do you need to subtract?

Answer 1

[Walter Roberson]

hint:

t = round([rand; rand; rand])
sum(t)

Execute that code several times in a row and look at the patterns you get

Answer 2

[Roger Stafford]

Let M be your 7 x 4 matrix of 1's and 0's.

 [m,n] = size(M);
 p = ceil(m*rand(1,n)) + m*(0:n-1);
 M(p) = 1-M(p);

[Samious]

So should I write this in matlab as function

      C = noisify(M)

      for M=[1 0 01; 0 0 1 0; 1 1 0 1; 0 0 1 1; 1 1 0 0; 0 1 1 1; 1 0 0 0]
      end
      size(M)=[m,n]
      p=ceil(m*rand(1,n))+m*(0:n-1);
      M(p) = 1-M(p);
      end?

[Walter Roberson]

No, there are several problems with that. For example, you should be working with the matrix that is passed in rather than with a fixed matrix.

Answer 3

[Image Analyst]

I shouldn't be doing this, but try this:

function test2()
    workspace;  % Make sure the workspace panel is showing.
    format longg;
    format compact;

      C = [1 0 0 1; 0 0 1 0; 1 1 0 1; 0 0 1 1; 1 1 0 0; 0 1 1 1; 1 0 0 0]
    % Make it noisy by flipping one bit in each column.
      noisyC = noisify(C)

function C = noisify(original_matrix)
  [rows, columns] = size(original_matrix)
  % Get the row in each column that you're going to invert.
  randomRowsToFlip = randi(rows, [columns, 1])
  % Initialize
  C = logical(original_matrix);
  % Now flip the appropriate row in each column.
  for col = 1 : columns
    C(randomRowsToFlip(col), col) = ~C(randomRowsToFlip(col), col);
  end

Now you can't turn it in, or else I'd get the A. Bad Image Analyst - I've slapped my hand as punishment.