From 81dca0351dd23ac090651e73b6f562a8bdc4e3da Mon Sep 17 00:00:00 2001
From: bthirion <bertrand.thirion@inria.fr>
Date: Tue, 24 Jan 2012 23:34:23 +0100
Subject: [PATCH] ENH: Some cleaning in the examples
---
examples/mixture/gmm_model_selection.py | 12 ++++++++++--
examples/mixture/plot_gmm_classifier.py | 6 +++---
2 files changed, 13 insertions(+), 5 deletions(-)
diff --git a/examples/mixture/gmm_model_selection.py b/examples/mixture/gmm_model_selection.py
index 8ea0ec61f2..cae9e3d06f 100644
--- a/examples/mixture/gmm_model_selection.py
+++ b/examples/mixture/gmm_model_selection.py
@@ -10,7 +10,12 @@ and the number of components in the model.
In that case, AIC also provides the right result (not shown to save time),
but BIC is better suited if the problem is to identify the right model.
Unlike Bayesian procedures, such inferences are prior-free.
+
+In that case, the full model with 2 componnets
+(which corresponds to the true generative model) is selected.
"""
+print __doc__
+
import itertools
import numpy as np
@@ -32,7 +37,7 @@ X = np.r_[np.dot(np.random.randn(n_samples, 2), C),
lowest_bic = np.infty
bic = []
n_components_range = range(1, 7)
-cv_types = ['spherical', 'diag', 'tied', 'full']
+cv_types = ['spherical', 'tied', 'diag', 'full']
for cv_type in cv_types:
for n_components in n_components_range:
# Fit a mixture of gaussians with EM
@@ -58,6 +63,9 @@ for i, (cv_type, color) in enumerate(zip(cv_types, color_iter)):
pl.xticks(n_components_range)
pl.ylim([bic.min() * 1.01 - .01 * bic.max(), bic.max()])
pl.title('BIC score per model')
+xpos = np.mod(bic.argmin(), len(n_components_range)) + .65 +\
+ .2 * np.floor(bic.argmin() / len(n_components_range))
+pl.text(xpos, bic.min() * 0.97 + .03 * bic.max(), '*', fontsize=14)
spl.set_xlabel('Number of components')
spl.legend([b[0] for b in bars], cv_types)
@@ -84,6 +92,6 @@ pl.xlim(-10, 10)
pl.ylim(-3, 6)
pl.xticks(())
pl.yticks(())
-pl.title('Selected GMM')
+pl.title('Selected GMM: full model, 2 components')
pl.subplots_adjust(hspace=.35, bottom=.02)
pl.show()
diff --git a/examples/mixture/plot_gmm_classifier.py b/examples/mixture/plot_gmm_classifier.py
index 835df829f6..b7bb63d37f 100644
--- a/examples/mixture/plot_gmm_classifier.py
+++ b/examples/mixture/plot_gmm_classifier.py
@@ -42,7 +42,7 @@ def make_ellipses(gmm, ax):
angle = np.arctan2(u[1], u[0])
angle = 180 * angle / np.pi # convert to degrees
v *= 9
- ell = mpl.patches.Ellipse(gmm._get_means()[n, :2], v[0], v[1],
+ ell = mpl.patches.Ellipse(gmm._get_means()[n, :2], v[0], v[1],
180 + angle, color=color)
ell.set_clip_box(ax.bbox)
ell.set_alpha(0.5)
@@ -78,9 +78,9 @@ pl.subplots_adjust(bottom=.01, top=0.95, hspace=.15, wspace=.05,
for index, (name, classifier) in enumerate(classifiers.iteritems()):
# Since we have class labels for the training data, we can
# initialize the GMM parameters in a supervised manner.
- classifier.means_ = np.array([X_train[y_train == i, :].mean(axis=0)
+ classifier.means_ = np.array([X_train[y_train == i].mean(axis=0)
for i in xrange(n_classes)])
-
+
# Train the other parameters using the EM algorithm.
classifier.fit(X_train, init_params='wc', n_iter=20)
--
GitLab