def surface_normals(S, number_of_neighbors):
nbrs = _NearestNeighbors(n_neighbors=number_of_neighbors,
algorithm='kd_tree').fit(S)
distances, indices = nbrs.kneighbors(S)
pca = _PCA()
normals_S = [] #np.zeros(2)
principle_axes = [] #np.zeros([2,2])
surface_region = _np.zeros([1, S.shape[1]])
for i in range(indices.shape[0]):
for j in indices[i]:
surface_region = _np.append(surface_region, S[j])
surface_region = _np.reshape(surface_region, (-1, S.shape[1]))
surface_region = _np.delete(surface_region, 0, 0)
pca_region = pca.fit(surface_region)
principle_axes.append(pca_region.components_)
#principle_axes = _np.append(principle_axes,pca_region.components_)
#normals_S = np.append(normals_S,pca_region.components_[1])
normals_S.append(pca_region.components_[1])
#normals_S = np.reshape(normals_S,(-1,2))
#principle_axes = np.reshape(principle_axes,(-1,2,2))
#return [np.delete(normals_S,0,0),principle_axes,distances]
return normals_S, principle_axes, distances
def with_pca(features, n_pcs=20, thresh=3):
"""PCA-based outlier removal function
Args:
features - the input feature data for finding outliers
n_pcs - the number of principal components to look for outliers in
thresh - the standard-deviation multiplier for outlier id
(e.g. thresh = 3 means values > 3 stdv from the mean will
be flagged as outliers)
Returns:
mask - a boolean array with True for good values, False for outliers
"""
# create the bool mask where outlier samples will be flagged as False
mask = _np.repeat(True, features.shape[0])
# set up the pca reducer, then transform the data
reducer = _PCA(n_components=n_pcs, whiten=True)
transformed = reducer.fit_transform(features)
# loop through the number of pcs set and flag values outside the threshold
for i in range(n_pcs):
outliers = abs(transformed[:, i]) > thresh
mask[outliers] = False
return mask
def __init__(self, options):
self.handle_options(options)
out_params = convert_params(options.get('params', {}),
ints=['k'],
aliases={'k': 'n_components'})
self.estimator = _PCA(**out_params)
def get_pca(self): #gets a PCA of the data the object was initialized with
"""
Principal component analysis of data
"""
pca = _PCA()
self.pca = pca.fit_transform(self._data)
self.pca_explained_var = pca.explained_variance_ratio_ * 100 #output a percent of the variance explained within the object defined here
return
def get_pca(self):
"""
Principal component analysis of data
"""
pca = _PCA()
self.pca = pca.fit_transform(self._data)
self.pca_explained_var = pca.explained_variance_ratio_ * 100
return
def _fit(self, X):
self.estimator_ = _PCA(iterated_power=self.iterated_power,
n_components=self.n_components,
random_state=self.random_state,
svd_solver=self.svd_solver,
tol=self.tol,
whiten=self.whiten).fit(X)
return self
def pca(features, n_pcs=100):
"""PCA transformation function
Args:
features - the input feature data to transform
n_pcs - the number of components to keep after transformation
Returns:
an array of PCA-transformed features
"""
reducer = _PCA(n_components=n_pcs, whiten=True)
return reducer.fit_transform(features)
def main():
print("Tesing the performance of PCA...")
# Generate data
seed(100)
X = randint(0, 10, (6, 3))
X_train, X_test = train_test_split(X)
# Fit and transform data.
pca = PCA()
pca.fit(X, n_components=2)
print("Imylu PCA:\n", pca.transform(X))
# Compare result with scikit-learn.
_pca = _PCA(n_components=2)
print("Sklearn PCA:\n", _pca.fit(X).transform(X))
def myPca(features, n_pcs=100):
reducer = _PCA(whiten=True)
#features = StandardScaler().fit_transform(features)
transformed = reducer.fit_transform(features)
# plt.figure()
# plt.plot(np.cumsum(reducer.explained_variance_ratio_))
# plt.xlabel('Number of Components')
# plt.ylabel('Variance (%)') #for each component
# plt.title('Modified Stanford-CCB Explained Variance')
# plt.grid()
# plt.show()
return reducer, transformed[:, 0:n_pcs]
def __init__(self, *args, **params):
try:
self.variables = args[:]
assert len(self.variables) > 0
except:
raise Exception('Syntax error: Expected "<field> ..."')
out_params = self.convert_params(
params,
ints=['k'],
aliases = {
'k': 'n_components'
}
)
self.estimator = _PCA(**out_params)
def __init__(self, **kwargs):
super().__init__()
self._method = _PCA(**kwargs)
self.hyperparams = self._method.get_params