Generating multivariate normal distribution with plot

multivariate.jl

# 2D multivariate normal
# with covariance matrix [3 2; 2 5], i.e. Var(x)=3, Var(y)=5, Cov(x,y)=2
# Break down covariacne matrix with Cholesky decomposition and multiply with iid std normal random numbers

Sigma = [4 9; 9 25]
S = chol(Sigma) # upper triangular part, i.e. S' * S == Sigma
iidnorm = randn(2,1000)
rv = S' * iidnorm # 2x1000 matrix of 1000 random number pairs conforms to the covariance matrix Sigma

# Plot in 2D
using Gadfly
plot(x=mvnorm[1,:], y=mvnorm[2,:], Geom.point)