how I write independent random numbers from a normal distribution, for example between (0,1)?

How I write independent random numbers from a normal distribution, for example between (0,1)? And how I can put this numbers in a matrix. Thank you.

Answer 1

[Massimo Zanetti]

There is specific matlab function https://it.mathworks.com/help/matlab/ref/randn.html

For example, create 4x5 matrix of normal distributed random numbers:

A=randn(4,5)

Notice that a Gaussian distributed random variable takes values between (-inf,inf), there is no Gaussian variable in (0,1).

randn assumes the distribution is Gaussian with mean=0 and variance=1. If you want to customize it, you can do as follows:

MU=2; %mean
SIG=5; %variance
A= SIG*randn(4,5)+MU;

[mike papas]

η with index k are independent random numbers from a normal distribution (written in maths η with index k ⇠ N(0,1),

[Walter Roberson]

nu(k) = randn();

If you want to create a vector of length N then

nu = randn(1,N);

The 1, N is the size, not the mean and standard deviation. For other means or standard deviations use the SIG MU transformation that Massimo Zanetti shows.

[mike papas]

Thank you very much! It was really helpful.

Answer 2

[Walter Roberson]

If you restrict values drawn from a Normal distribution to any particular range (such as 0 to 1), then the values will no longer be normally distributed. It is impossible for anything to be Normally distributed unless it has infinite tails in both direction. MATLAB's randn() can produce values from -realmax to +realmax which is not truly infinite but the probability lost is tiny. Restricting to (0, 1) would be a significant probability loss and a major distortion of the normal distribution.

If you need non-uniform random numbers that must be within a particular finite range, then you should be considering using the Beta distribution; https://www.mathworks.com/help/stats/beta-distribution-1.html and http://www.mathworks.com/help/stats/betapdf.html