def clustering_gmm_modular(fm_train=generated, n=2, min_cov=1e-9, max_iter=1000, min_change=1e-9, cov_type=0): from shogun.Distribution import GMM from shogun.Features import RealFeatures from shogun.Library import Math_init_random Math_init_random(5) feat_train = RealFeatures(generated) est_gmm = GMM(n, cov_type) est_gmm.train(feat_train) est_gmm.train_em(min_cov, max_iter, min_change) return est_gmm
def clustering_gmm_modular (fm_train=generated,n=2,min_cov=1e-9,max_iter=1000,min_change=1e-9,cov_type=0): from shogun.Distribution import GMM from shogun.Features import RealFeatures from shogun.Library import Math_init_random Math_init_random(5) feat_train=RealFeatures(generated) est_gmm=GMM(n, cov_type) est_gmm.train(feat_train) est_gmm.train_em(min_cov, max_iter, min_change) return est_gmm
min_change) est_em_gmm = GMM(3, cov_type) est_em_gmm.train(feat_train) est_em_gmm.set_nth_mean(array([2.0, 0.0]), 0) est_em_gmm.set_nth_mean(array([-2.0, -2.0]), 1) est_em_gmm.set_nth_mean(array([-3.0, -3.0]), 2) est_em_gmm.set_nth_cov(array([[1.0, 0.0], [0.0, 1.0]]), 0) est_em_gmm.set_nth_cov(array([[1.0, 0.0], [0.0, 1.0]]), 1) est_em_gmm.set_nth_cov(array([[1.0, 0.0], [0.0, 1.0]]), 2) est_em_gmm.set_coef(array([0.3333, 0.3333, 0.3334])) print est_em_gmm.train_em(min_cov, max_em_iter, min_change) min_x_gen = min(min(generated[[0]])) - 0.1 max_x_gen = max(max(generated[[0]])) + 0.1 min_y_gen = min(min(generated[[1]])) - 0.1 max_y_gen = max(max(generated[[1]])) + 0.1 plot_real = empty(0) plot_est_smem = empty(0) plot_est_em = empty(0) for i in arange(min_x_gen, max_x_gen, 0.05): for j in arange(min_y_gen, max_y_gen, 0.05): plot_real = append(plot_real, array([real_gmm.cluster(array([i, j]))[3]])) plot_est_smem = append(plot_est_smem,
real_gmm.set_nth_mean(array([2.0]), 2) real_gmm.set_nth_cov(array([[0.3]]), 0) real_gmm.set_nth_cov(array([[0.1]]), 1) real_gmm.set_nth_cov(array([[0.2]]), 2) real_gmm.set_coef(array([0.3, 0.5, 0.2])) generated=array([real_gmm.sample()]) for i in range(199): generated=append(generated, array([real_gmm.sample()]), axis=1) feat_train=RealFeatures(generated) est_gmm=GMM(3) est_gmm.train(feat_train) est_gmm.train_em(min_cov, max_iter, min_change) est_mean1=est_gmm.get_nth_mean(0) est_mean2=est_gmm.get_nth_mean(1) est_mean3=est_gmm.get_nth_mean(2) est_cov1=est_gmm.get_nth_cov(0) est_cov2=est_gmm.get_nth_cov(1) est_cov3=est_gmm.get_nth_cov(2) est_coef=est_gmm.get_coef() print est_mean1 print est_cov1 print est_mean2 print est_cov2 print est_mean3 print est_cov3 print est_coef
real_gmm.set_nth_cov(array([[0.1]]), 1) real_gmm.set_nth_cov(array([[0.2]]), 2) real_gmm.set_coef(array([0.3, 0.5, 0.2])) generated=array([real_gmm.sample()]) for i in range(199): generated=append(generated, array([real_gmm.sample()]), axis=1) feat_train=RealFeatures(generated) est_smem_gmm=GMM(3) est_smem_gmm.train(feat_train) print est_smem_gmm.train_smem(max_iter, max_cand, min_cov, max_em_iter, min_change) est_em_gmm=GMM(3) est_em_gmm.train(feat_train) print est_em_gmm.train_em(min_cov, max_em_iter, min_change) min_gen=min(min(generated)) max_gen=max(max(generated)) plot_real=empty(0) plot_est_smem=empty(0) plot_est_em=empty(0) for i in arange(min_gen, max_gen, 0.001): plot_real=append(plot_real, array([real_gmm.cluster(array([i]))[3]])) plot_est_smem=append(plot_est_smem, array([est_smem_gmm.cluster(array([i]))[3]])) plot_est_em=append(plot_est_em, array([est_em_gmm.cluster(array([i]))[3]])) real_plot=plot(arange(min_gen, max_gen, 0.001), plot_real, "b") est_em_plot=plot(arange(min_gen, max_gen, 0.001), plot_est_em, "g") est_smem_plot=plot(arange(min_gen, max_gen, 0.001), plot_est_smem, "r") real_hist=hist(generated.transpose(), bins=50, normed=True, fc="gray") legend(("Real GMM", "Estimated EM GMM", "Estimated SMEM GMM"))
import numpy as np import os from shogun.Features import RealFeatures from shogun.Distribution import GMM from shogun.Library import Math_init_random # Load the data. f = open(os.path.dirname(__file__) + '../data/mvnrnd.data') data = np.fromfile(f, dtype=np.float64, sep=' ') data = data.reshape(-1, 2) f.close() Math_init_random(5) feat = RealFeatures(data.T) # Calculate mixture of Gaussians. gmm = GMM(2, 0) gmm.set_features(feat) gmm.train_em() # Vector of covariances; one for each Gaussian. print gmm.get_nth_cov(0) print gmm.get_nth_cov(1) # The vector of means. print gmm.get_nth_mean(0) print gmm.get_nth_mean(1) # The a priori weights of each Gaussian. print gmm.get_coef()