Generate Gaussian Mixture Distribution samples for PEM

How to generate samples from GMM distribution? I wanna set them to point estimation method.

Answer 1

[Bruno Luong]

% Characteristics of your GM distribution
mu = [1,2,5]
sigma = [0.2, 0.3, 0.4];
A = [5 3 2];

% number of samples
n = 10000;

c = [0, cumsum(A(:)')];
c = c/c(end);
[~,i] = histc(rand(1,n),c);
r = randn(1,n).*sigma(i) + mu(i);

hist(r,100)

[Bruno Luong]

Depending on your definition you might multiply A by sigma before feed it to the cumsum command.

c = [0, cumsum((A(:).*sigma(:))')];

[Bahman]

Thank you Bruno Luong. But how can estimate 2 point?

[Bruno Luong]

Sorry I don't know point estimation method to answer.