bay_gauss_pca = BayesianGaussianMixture(n_components=8, covariance_type='full', init_params='kmeans', weight_concentration_prior_type='dirichlet_process', weight_concentration_prior=0.1, random_state=SEED)