Generating random streams marginally different

Is there anyway in Matlab to generate streams of random normal numbers that are marginally different? By that I mean if the first stream would look like:

 [-0.9625    0.8736    0.8499    1.6579    0.6706   -2.3623   -0.7797   -0.5591    0.8103    2.9868]

then the next stream would be similar to the first stream, maybe something like:

 [-0.6321    0.9430    0.7002    1.3456    0.2832   -2.0011   -1.0188   -0.5932    1.1120    2.8932]

and so on with the next streams (each stream be similar to the other ones already generated).

Is there already any tools in Matlab doing this?

Thank you,

Kian

[Andrew Newell]

Perhaps a random walk might do it - depending on what you wanted the streams for.

Answer 1

[John D'Errico]

A statistician might say you wanted to generate a pair of correlated random variables. mvnrnd would help here.

Lacking that, you could just do this...

A = randn(1,10);
B = A + randn(size(A))/10;

Note that this is equivalent to a specific covariance matrix, so a similar stream could be gained from mvnrnd.

[Joep]

I assume you meant this.

    A = randn(1,10);
    B = A + randn(size(A))/10;

[John D'Errico]

Yes. It was just a shifty shift key at fault.

[Kian]

Thank you for your answer. With your answer I realized I had mixed up something.