Multivariate skew normal distribution

Hi,

Since version 2.30 the multivariate Student-t distribution (Cholesky parameterization) is available as stated in the docs (here and here).
When can we expect the multivariate skew normal distribution? Or is someone willing to share his custom function / implementation of this?

  real multi_skew_normal_lpdf(array[] vector y, vector mu, vector alpha, matrix L) {
    int N = dims(y)[1];
    vector[N] alpha_y; 
    
    for (n in 1:N) {
      alpha_y[n] = dot_product(alpha, y[n] - mu);
    }
    
    return N * log2() + multi_normal_cholesky_lpdf(y | mu, L) + std_normal_lcdf(alpha_y);
  }
5 Likes