could anyone help me how to generate two different random numbers of same value.

I am having two matrices A and B generated randomly with different values

A=rand(4,3)
B=rand(4,3)

Now I want to know is there any other way such that the some of the values generated in A can be made equal to the values generated in B.

[Yongjian Feng]

Not very sure what you need here. Since rand is random, it is possible already some of the values in A can be repeated in B.

1. Do you want the sum of A is the same as the sum of B?
2. Do you want exactly 3 (for examples) values in A to be already repeated in B?

[jaah navi]

Yes, I know that some of the values in A can be repeated in B.

In my case, the size of A and B is very large i.e, (48000,1). When I generate i can find only minimum number of values remains to be same i.e, fo example 2300 out of 48000. But I want the maximum number of values to be same i.e, 45700/48000.

Could you please hlp me onthis.

Answer 1

[Steven Lord]

A = rand(5, 5);
A1 = A; A2 = A; % Make copies for later comparison
B = rand(5, 5);
numberOfMatches = 6;
locationOfMatchesInB = randperm(numel(B), numberOfMatches);

If you want the elements from B to be in exactly the same locations in A:

A1(locationOfMatchesInB) = B(locationOfMatchesInB);

If you want the elements from B to be put in some location, but not necessarily the same locations, in A:

locationOfMatchesInA = randperm(numel(A), numberOfMatches);
A2(locationOfMatchesInA) = B(locationOfMatchesInB);

Now compare:

A

A = 5×5

0.9605 0.1936 0.1763 0.7655 0.5840 0.6432 0.6447 0.2002 0.5192 0.1195 0.8839 0.0211 0.4184 0.5463 0.9373 0.2247 0.7324 0.8192 0.9718 0.4654 0.3539 0.5847 0.8463 0.1017 0.9999

A1

A1 = 5×5

0.3280 0.1936 0.1763 0.7655 0.7112 0.6432 0.6447 0.2002 0.5192 0.7247 0.8839 0.9045 0.3648 0.5463 0.9373 0.2247 0.7324 0.8192 0.9718 0.4654 0.4195 0.5847 0.8463 0.1017 0.9999

A2

A2 = 5×5

0.9605 0.1936 0.1763 0.7655 0.5840 0.7112 0.6447 0.2002 0.5192 0.1195 0.8839 0.7247 0.4184 0.5463 0.9373 0.2247 0.9045 0.3280 0.9718 0.4195 0.3648 0.5847 0.8463 0.1017 0.9999

B

B = 5×5

0.3280 0.3758 0.7772 0.0375 0.7112 0.4655 0.5999 0.1197 0.2314 0.7247 0.4005 0.9045 0.3648 0.0259 0.9559 0.6848 0.5505 0.0491 0.4264 0.4717 0.4195 0.3723 0.9590 0.2997 0.9910

Where did A1 flip?

[sort(locationOfMatchesInB).', find(A1 ~= A)]

ans = 6×2

1 1 5 5 8 8 13 13 21 21 22 22

What changed?

A1(A1 ~= A)

ans = 6×1

0.3280 0.4195 0.9045 0.3648 0.7112 0.7247

B(A1 ~= A)

ans = 6×1

0.3280 0.4195 0.9045 0.3648 0.7112 0.7247

[jaah navi]

For the code,

A = rand(5, 5);
A1 = A; A2 = A; % Make copies for later comparison
B = rand(5, 5);
numberOfMatches = 6;
locationOfMatchesInB = randperm(numel(B), numberOfMatches);

Is there any way to make it invisible in the code

numberOfMatches

[Walter Roberson]

A = rand(5, 5);
A1 = A; A2 = A; % Make copies for later comparison
B = rand(5, 5);
locationOfMatchesInB = randperm(numel(B), 6);

Answer 2

[Amit Bhowmick]

Check the values you will alyas get zero that is they are not equal. More preciesly probability of getting same value is very less. Basically the algorithm of generating random number is such you can learn here

https://en.wikipedia.org/wiki/Random_number_generation

rand(4,3)==rand(4,3) or isequal(rand(4,3),rand(4,3))

Answer 3

[Walter Roberson]

A = sort(rand(300,300));
imagesc(A)
title('original')

p = 45700/48000;
mask = rand(size(A)) <= p;
B = [A(mask); rand(numel(A)-nnz(mask),1)];
B = reshape(B(randperm(numel(B))), size(A));
imagesc(B)
title('rerandom')

mean(ismember(B(:), A)) * 100

ans = 95.2233

p * 100

ans = 95.2083

So in this case, the target fraction to keep was 95.2083% and the actual fraction kept was 95.2233%

If it was necessary to have exactly the ratio 45700/48000 stay the same then that could be done... though it becomes trickier if you are working with integers.

[jaah navi]

These are the two arrays A and B

A= arrayfun( @my_rand, repelem((1:10).',30), 'UniformOutput', false)
B= arrayfun( @my_rand, repelem((1:10).',30), 'UniformOutput', false)

What I actually need is I want B to have atleast 270 values which is similar to A.

Could you please help me on this.

[Walter Roberson]

A = arrayfun( @my_rand, repelem((1:10).',30), 'UniformOutput', false);
B = arrayfun( @my_rand, repelem((1:10).',30), 'UniformOutput', false);

p = 0.9;
nkeep = ceil(p * numel(A)) ;
keepidx = randperm(numel(A), nkeep); 
B(keepidx) = A(keepidx);

[jaah navi]

thanks for your help.