pub fn generate_samples(
nk: &[usize],
d: usize
) -> (Array2<f64>, Array2<f64>, Array2<f64>, Array3<f64>)
Expand description
Create data generated with a Gaussian mixture model. Returns $n_1+\ldots +n_k$ samples from a Gaussian mixture with $k$ components in a $d$-dimensional feature space. It also returns the $(n_1+\ldots +n_k) \times k$ “true” responisiblity matrix (i.e., only ones and zeros in its elements). For testing, it returns also the generated covariances Returns: (samples, responsibilities, means, covariances)