writing normally distributed error code

Hello all,

I have an equation

Y = aX + E.....

E is error term that is N ~ ( 0 , [ 25 0 ; 0 25] ) , how I can write a code for E in this case know that it will be 2 x1 matrix size. that is normally distibuted with 0 mean and 25 standard deviation

Thank you in advance.

[Lina Alhelo]

You mean it will be as random value, I will not have a certain constant value?

[Bjorn Gustavsson]

Yes, that is what N(m,C) means, that you have a random numbers with expected values m and covariance C. Your case is what one typically do to model normal-distributed "random noise in measurements".

[Lina Alhelo]

Thank you very much.

[Bjorn Gustavsson]

I typically use something like this for scalar white noise with standard deviation proportional to the sqrt(intensity):

I = I_ideal + I_ideal.^.5*randn(size(I_ideal));

Maybe the documentation to normrnd gives you more information?

HTH

[Lina Alhelo]

Thank you very much, your help was valuable.

Answer 1

[Bjorn Gustavsson]

Have a look at: randn, and normrnd.

HTH