generating a random variable with a given mean and matrix

can any one tell me how to generate a NORMAL DISTRIBUTION random variable X = N(mu,A), where mu is the mean (mu = 0, 0 mean) and A is a p*p IDENTITY MATRIX. Please ask me if the question is unclear. Thanks

[Diego Jaramillo]

mvnrnd

Answer 1

[Image Analyst]

Like this?

numRows = size(A, 1);
A(logical(eye(numRows))) = randn(1, numRows)

[fafz1203]

may I ask what the first line of command does?

[fafz1203]

in my particular case A is 10*10 identity matrix

[Image Analyst]

The first line just figures out how large your square A matrix is. The second line assigns random numbers to elements along the diagonal only.

[fafz1203]

that means i do not need the first line if i already know the size of my matrix right?

[Image Analyst]

Correct. If you already know numRows, or whatever you call your variable, if you even use one, then you do not need to calculate it again. You can do this:

A(logical(eye(10))) = randn(1, 10)

[fafz1203]

thanks

[fafz1203]

I'm sorry but it turns out that I was wrong, Here is what I want to do I want to generate a random variable X (X~N(0,A)), that is it has 0 mean and A is a 10*10 covariance matrix which is a 10*10 identity matrix

now what i want to do is I want to generate a 10*1000 random variable X using the above information.

Please help me if you can. Thanks

[Image Analyst]

I don't believe you can do that and have the random variable be random. If it's random, you won't have A=eye(10) for the covariance matrix.