def construct_pca_all_comp(self):
pca = plinear.PCA(
standardize=self.standardize,
max_components=0,
variance_covered=1,
use_generalized_eigenvectors=self.use_generalized_eigenvectors)
return pca
def construct_pca(self):
max_components = self.max_components
variance_covered = self.variance_covered
pca = plinear.PCA(
standardize=self.standardize,
max_components=max_components,
variance_covered=variance_covered / 100.0,
use_generalized_eigenvectors=self.use_generalized_eigenvectors)
return pca
def test_can_pickle_and_unpickle(self):
self.create_normal_dataset()
projector = linear.PCA(variance_covered=.99)(self.dataset)
pickled = pickle.dumps(projector)
restored = pickle.loads(pickled)
self.assertFalse((projector.projection - restored.projection).any())
self.assertFalse((projector.center - restored.center).any())
self.assertFalse((projector.scale - restored.scale).any())
transformed, new_transformed = projector(self.dataset), restored(self.dataset)
print transformed[0][0]
for ex1, ex2 in zip(transformed, new_transformed):
for v1, v2 in zip(ex1, ex2):
self.assertEqual(v1, v2)