def test_PCA(self):
"""
Regression test.
"""
trajectory_handler = TrajectoryHandlerStub(testPCAMetric.not_iterposed_coordsets,66)
clustering = Clustering([Cluster(None,range(6)),Cluster(None,range(6,12))], "a clustering")
pcaMetric = PCAMetric(trajectory_handler)
self.assertAlmostEquals(pcaMetric.evaluate(clustering), 1.427748687873, 12)
def test_eigenvalues(self):
# do it with PCA metric
my_cov_matrix = PCAMetric.create_covariance_matrix(testPCAMetric.coordsets)
biggest_eigenvalue = PCAMetric.calculate_biggest_eigenvalue(my_cov_matrix)
# Do it with prody
pca = prody.PCA('pcametric_pca')
pca.buildCovariance(testPCAMetric.ensemble)
pca.calcModes(n_modes=1)
self.assertAlmostEqual(pca.getEigvals()[0], biggest_eigenvalue,10)
def test_PCA(self):
"""
Regression test.
"""
trajectory_handler = TrajectoryHandlerStub(
testPCAMetric.not_iterposed_coordsets, 66)
clustering = Clustering(
[Cluster(None, range(6)),
Cluster(None, range(6, 12))], "a clustering")
pcaMetric = PCAMetric(trajectory_handler)
self.assertAlmostEquals(pcaMetric.evaluate(clustering), 1.427748687873,
12)
def test_eigenvalues(self):
# do it with PCA metric
my_cov_matrix = PCAMetric.create_covariance_matrix(
testPCAMetric.coordsets)
biggest_eigenvalue = PCAMetric.calculate_biggest_eigenvalue(
my_cov_matrix)
# Do it with prody
pca = prody.PCA('pcametric_pca')
pca.buildCovariance(testPCAMetric.ensemble)
pca.calcModes(n_modes=1)
self.assertAlmostEqual(pca.getEigvals()[0], biggest_eigenvalue, 10)
def test_covariance_matrix_vs_prody(self):
# do it with PCA metric
my_cov_matrix = PCAMetric.create_covariance_matrix(testPCAMetric.coordsets)
# Do it with prody
pca = prody.PCA('pcametric_pca')
pca.buildCovariance(testPCAMetric.ensemble)
prody_cov_matrix = pca._cov
# Compare
numpy.testing.assert_almost_equal(my_cov_matrix, prody_cov_matrix,10)
def test_covariance_matrix_vs_prody(self):
# do it with PCA metric
my_cov_matrix = PCAMetric.create_covariance_matrix(
testPCAMetric.coordsets)
# Do it with prody
pca = prody.PCA('pcametric_pca')
pca.buildCovariance(testPCAMetric.ensemble)
prody_cov_matrix = pca._cov
# Compare
numpy.testing.assert_almost_equal(my_cov_matrix, prody_cov_matrix, 10)
def analysis_function_pca(self, clustering, trajectory_handler):
calculator = PCAMetric(trajectory_handler)
return calculator.evaluate(clustering)
def analysis_function_pca(self, clustering, data_handler):
"""
"""
calculator = PCAMetric(data_handler)
return calculator.evaluate(clustering)
def analysis_function_pca(self,clustering, data_handler):
"""
"""
calculator = PCAMetric(data_handler)
return calculator.evaluate(clustering)