randn

Pseudo-random number following the standard normal distribution

Usages

randn

• It returns a pseudo-random number following the standard normal distribution.

randn(n)

• It returns an n-by-n matrix of pseudo-random numbers following the standard normal distribution.
• n should be a real, non-negative integer scalar. If n is zero, it returns an empty array.

randn(s1,s2,...,sN)

• It returns an array of size [s1, s2, ..., sN], which contains pseudo-random numbers following the standard normal distribution.
• s1, s2, ..., sN should be real, non-negative integer scalars. If one of s1, s2, ..., sN is zero, it returns an empty array.

randn(sizeVec)

• It returns an array of pseudo-random numbers, which follows the standard normal distribution. The array's size is specified by the size vector sizeVec.
• sizeVec should contain real, non-negative elements. If an element of sizeVec is zero, it returns an empty array.

Examples

• In this example, we scale and shift 100,000 samples generated by randn (in Line 2). Then, we estimate the mean and standard deviation from the samples. We find that $\mu\approx 1$ and $\sigma\approx 2$.

Input

% Vector of numbers from normal distribution with mean equal to 1 and SD equal to 2.
r = 2 * randn(1,100000) + 1;
% Mean estimated from r.
mean(r)
% Standard deviation estimated from r.
sqrt(mean((1-r).^2))

Output

ans = 1e-1 × 
 9.9262

ans = 
 2.0067

Discussion

• Since this app's algorithms for generating random numbers and initial states may be different, users should not expect this app and MATLAB to give the same sequences of pseudo-random numbers.