def fit(self, X, y):
#creating a manifold on training data
self.model = LocallyLinearEmbedding(
method=self.method,
n_neighbors=self.n_neighbors,
n_components=self.n_components,
reg=self.reg,
eigen_solver=self.eigen_solver,
random_state=self.random_state).fit(X, y)
#determining centroids for given points
self.centroids = KMeans(n_clusters=self.n_clusters,
random_state=self.random_state).fit(
self.model.transform(X))
labels = self.centroids.predict(self.model.transform(
X)) # Every point is assigned to a certain cluster.
#assigning each centroid to the correct cluster
confusion_m = confusion_matrix(y, labels)
m = Munkres()
cost_m = make_cost_matrix(confusion_m)
target_cluster = m.compute(
cost_m) # (target, cluster) assignment pairs.
#saving mapping for predictions
self.mapping = {
cluster: target
for target, cluster in dict(target_cluster).items()
}
def get_dim_reds_scikit(pct_features):
n_components = max(int(pct_features * num_features), 1)
return [
LinearDiscriminantAnalysis(n_components=n_components),
TruncatedSVD(n_components=n_components),
#SparseCoder(n_components=n_components),
DictionaryLearning(n_components=n_components),
FactorAnalysis(n_components=n_components),
SparsePCA(n_components=n_components),
NMF(n_components=n_components),
PCA(n_components=n_components),
RandomizedPCA(n_components=n_components),
KernelPCA(kernel="linear", n_components=n_components),
KernelPCA(kernel="poly", n_components=n_components),
KernelPCA(kernel="rbf", n_components=n_components),
KernelPCA(kernel="sigmoid", n_components=n_components),
KernelPCA(kernel="cosine", n_components=n_components),
Isomap(n_components=n_components),
LocallyLinearEmbedding(n_components=n_components,
eigen_solver='auto',
method='standard'),
LocallyLinearEmbedding(n_neighbors=n_components,
n_components=n_components,
eigen_solver='auto',
method='modified'),
LocallyLinearEmbedding(n_neighbors=n_components,
n_components=n_components,
eigen_solver='auto',
method='ltsa'),
SpectralEmbedding(n_components=n_components)
]
def get_metastable_connections_from_gmm(
data,
gmm,
connection_estimation_method='max_path_distance_diff',
min_paths=3,
distance='euclidean',
low_dimension_distances=True,
as_graph=False):
means = gmm.means_
memberships = gmm.predict(data)
if connection_estimation_method in [
'max_path_distance_diff', 'connecting_paths', 'mst'
]:
if low_dimension_distances:
pca = PCA(n_components=2)
lle = LocallyLinearEmbedding(n_components=2,
n_neighbors=int(0.8 * data.shape[0]))
distance_matrix = squareform(
pdist(lle.fit_transform(data), distance))
else:
distance_matrix = squareform(pdist(data, distance))
weighted_graph = nx.Graph(distance_matrix)
else:
weighted_graph = None
return get_metastable_connections(data,
means,
memberships,
method=connection_estimation_method,
weighted_graph=weighted_graph,
min_paths=3,
as_graph=as_graph)
def function(self, data):
# pylint: disable=not-a-mapping
lle = LocallyLinearEmbedding(n_neighbors=self.n_neighbors,
n_components=self.n_components,
**self.kwargs)
emb = lle.fit_transform(data)
return emb
def get_lower_dimensional_projection(cluster_data,
algorithm='tsne',
projection_dim=2):
if algorithm.lower() == 'tsne':
tsne_object = TSNE(n_components=projection_dim, random_state=42)
lower_dimensional_projected_data = tsne_object.fit_transform(
cluster_data)
return lower_dimensional_projected_data
elif algorithm.lower() == 'pca':
pca_object = PCA(n_components=projection_dim,
random_state=42,
copy=False)
lower_dimensional_projected_data = pca_object.fit_transform(
cluster_data)
return lower_dimensional_projected_data
elif algorithm.lower() == "mds":
mds_object = MDS(n_components=projection_dim, random_state=42)
lower_dimensional_projected_data = mds_object.fit_transform(
cluster_data)
return lower_dimensional_projected_data
else:
lle_object = LocallyLinearEmbedding(n_components=projection_dim,
random_state=42)
lower_dimensional_projected_data = lle_object.fit_transform(
cluster_data)
return lower_dimensional_projected_data
def lle(space):
n_neighbors = int(space['n_neighbors'])
method = space['method']
vertices, colors = get_all_vertices_dk_atlas_w_colors()
print(space)
lle = LLE(n_neighbors=n_neighbors, n_components=2, method=method, neighbors_algorithm='auto')
lle_xy = lle.fit_transform(vertices)
centers = get_centers_of_rois_xy(lle_xy)
avg_distance = avg_distance_between_center_of_masses(centers)
model_name = 'lle_{}_{}'.format(method, avg_distance)
result = {
'loss': -avg_distance,
'space': space,
'status': STATUS_OK
}
save_json_result(model_name, result)
save_2d_roi_map(lle_xy, colors, centers, model_name)
return result
def applyLlleWithStandardisation(data, n_components=None):
X = preprocessing.scale(data)
lle = LocallyLinearEmbedding(n_components=n_components,
eigen_solver="auto")
return lle.fit_transform(X)
def evaluate_embeddings(D, labels):
estimators = [
KMeans(init='k-means++', n_clusters=5, n_init=10)
] #,AgglomerativeClustering(n_clusters=5),AgglomerativeClustering(n_clusters=5,linkage='average')]
est_names = [
'KMeans'
] #,'wardAgglomerativeClustering','avgAgglomerativeClustering']
for e in range(len(estimators)):
print '!!----------------------------------!!'
print est_names[e]
estim = estimators[e]
for i in range(2, 6 + 1):
print '--------------------------------------'
print '#dim = ' + str(i)
model_t = TSNE(n_components=i,
learning_rate=100,
perplexity=10,
method='exact')
x = model_t.fit_transform(D)
bench_k_means(estim, name="tsne", data=x, labels=labels)
model_i = Isomap(n_components=i)
x = model_i.fit_transform(D)
bench_k_means(estim, name="isomap", data=x, labels=labels)
model_l = LocallyLinearEmbedding(n_components=i)
x = model_l.fit_transform(D)
bench_k_means(estim, name="lle", data=x, labels=labels)
def finalData(algo,lb,dm):
Page3_Util.a.maindata.make_XY(lb)
pca = LocallyLinearEmbedding(n_components=dm)
if algo=='PCA':
pca = PCA(n_components=dm)
elif algo=='Linear Embading':
pca = LocallyLinearEmbedding(n_components=dm)
elif algo== 'Isomap':
pca = Isomap(n_components=dm)
elif algo== 'MDS':
pca = MDS(n_components=dm)
elif algo== 'SpectralEmbedding':
pca = SE(n_components=dm)
else:
if dm==Page3_Util.a.maindata.N_features:dm=dm-1
pca = TSNE(n_components=dm)
principalComponents = pca.fit_transform(Page3_Util.a.maindata.X)
principalDf = pd.DataFrame(data=principalComponents
, columns=["D{}".format(i) for i in range(dm)])
finalDf = pd.concat([principalDf, Page3_Util.a.maindata.df[[lb]]], axis=1)
csv_string = finalDf.to_csv(index=False, encoding='utf-8')
csv_string = "data:text/csv;charset=utf-8," + urllib.parse.quote(csv_string)
return csv_string
def initial_embed(self, reduce, d):
reduce = reduce.lower()
assert reduce in ['isomap', 'ltsa', 'mds', 'lle', 'se', 'pca', 'none']
if reduce == 'isomap':
from sklearn.manifold import Isomap
embed = Isomap(n_components=d)
elif reduce == 'ltsa':
from sklearn.manifold import LocallyLinearEmbedding
embed = LocallyLinearEmbedding(n_components=d,
n_neighbors=5,
method='ltsa')
elif reduce == 'mds':
from sklearn.manifold import MDS
embed = MDS(n_components=d, metric=False)
elif reduce == 'lle':
from sklearn.manifold import LocallyLinearEmbedding
embed = LocallyLinearEmbedding(n_components=d,
n_neighbors=5,
eigen_solver='dense')
elif reduce == 'se':
from sklearn.manifold import SpectralEmbedding
embed = SpectralEmbedding(n_components=d)
elif reduce == 'pca':
from sklearn.decomposition import PCA
embed = PCA(n_components=d)
if reduce == 'none':
self.embed = lambda x: x
else:
self.embed = lambda x: embed.fit_transform(x)
def LLE_plot(data):
"""
This function print and plots the result of LLE(Local Linear Embedding) algorithm.
"""
print("Computing LLE embedding")
t1 = time()
for n in range(1, 50):
plt.figure(figsize=(16,9))
n_neighbors = n
print("n_neighbors = %d"%n_neighbors)
for i in range(10):
condition = data['label'] == i
subset_data = data[condition]
clf = LocallyLinearEmbedding(n_neighbors, n_components=2, method='standard', eigen_solver='dense')
t0 = time()
X_lle = clf.fit_transform(subset_data)
print("Done. Reconstruction error: %g" % clf.reconstruction_error_)
print("Locally Linear Embedding of the digits (time %.2fs)" %(time() - t0))
plt.scatter(X_lle[:, 0], X_lle[:, 1], cmap=plt.cm.hot, s=2, label='digit %d'%i)
plt.ylim([-0.1, 0.1])
plt.xlim([-0.2, 0.2])
plt.legend()
plt.grid()
plt.savefig("./img/n-neighbor=%d.png"%n_neighbors, dpi=300)
print("totally consumed time : (%.2fs)" %(time() - t1))
def data_lle_preprocessing(data, feature_columns):
data = data.dropna()
sc = preprocessing.StandardScaler()
data[feature_columns] = sc.fit_transform(data[feature_columns])
lle = LocallyLinearEmbedding(n_components=4)
data[feature_columns[:-1]] = lle.fit_transform(data[feature_columns])
return data, feature_columns[:-1]
def wrap_lle(x, required_d, neighbors):
# 对输入x，用LLE方法降维到required_d维，并将降维后的数据保存为np文件，方便下次调用
lle = LocallyLinearEmbedding(n_components=required_d,
n_neighbors=neighbors)
lle.fit(x)
x_lle = lle.embedding_
np.save('LLE/np_x_LLE_' + str(required_d) + str(neighbors), x_lle)
return x_lle
def score_lle(x_train, y_train, x_test, y_test):
lle = LocallyLinearEmbedding(n_neighbors=5, n_components=4)
x_train = lle.fit_transform(x_train)
x_test = lle.fit_transform(x_test)
nb = GaussianNB()
nb.fit(x_train, y_train)
y_pred = nb.predict(x_test)
return accuracy_score(y_pred, y_test)
def lle(x):
"""
Useful link:
https://stackoverflow.com/questions/42275922/setting-the-parameters-of-locally-linear-embedding-lle-method-in-scikit-learn
"""
embedding = LocallyLinearEmbedding(n_components=2) # 2D projection
x_transformed = embedding.fit_transform(x)
return embedding, x_transformed
def locally_linear_embedding(self,
n_neighbors=5,
n_components=3,
reg=1e-3,
eigen_solver='auto',
tol=1e-6,
max_iter=100,
method='standard',
hessian_tol=1E-4,
modified_tol=1E-12,
neighbors_algorithm='auto',
random_state=None,
n_jobs=None):
"""Computes the locally linear embedding of x_data.
Args:
n_neighbors: An integer, which is the number of neighbors
considered for each point
n_components: An integer, which is the number of coordinates
for the manifold
reg: A float, which is the regularization constant
eigen_solver: A string ('auto', 'arpack', 'dense'),
which is solver for the problem
tol: A float, which is the convergence tolerance for
eigen solvers (arpack)
max_iter: An integer, which is the max number of iteration
for the arpack solver
method: A string ('standard', 'hessian', 'modified', 'ltsa'),
which is the embedding algorithm
hessian_tol: A float, which is the tolerance for Hessian method
modified_tol: A float, which is the tolerance for LLE method
neighbors_algorithm: A string ('auto', 'brute',
'kd_tree', 'ball_tree'), which is the
algorithm for nearest neighbors search
random_state: An integer, which is a seed for random number
generator
n_jobs: An integer (-1 all), which is the number of parallel
jobs to run
Returns:
A numpy ndarray, which has a shape like
(length of x_data, n_components)
"""
x_data = self.x_data.reshape(
(self.x_data.shape[0], np.prod(self.x_data.shape[1:])))
lle = LocallyLinearEmbedding(n_neighbors=n_neighbors,
n_components=n_components,
reg=reg,
eigen_solver=eigen_solver,
tol=tol,
max_iter=max_iter,
method=method,
hessian_tol=hessian_tol,
modified_tol=modified_tol,
neighbors_algorithm=neighbors_algorithm,
random_state=random_state,
n_jobs=n_jobs)
return lle.fit_transform(x_data)
def LLE_dr(data_set, com_dimentions):
print("Dimentions of Dataset before LLE: ", data_set.shape)
lle = LocallyLinearEmbedding(n_components=com_dimentions)
data_transform = lle.fit_transform(data_set) # Fit the model with X and Apply dimensionality reduction to X.
print("Dimentions of Dataset after LLE: ", data_transform.shape)
return data_transform
def KLLE(feature, dim):
print("dim:", dim)
t = time.time()
lle = LocallyLinearEmbedding(n_components=dim,
n_jobs=4,
neighbors_algorithm='ball_tree')
feature_ = lle.fit_transform(feature)
np.save('LLE/feature_' + str(dim), feature_)
print("time:", time.time() - t)
def fit_transform(self, X, n_components=2):
if self.method == 'pca':
self.compresser = PCA(n_components=n_components)
elif self.method == 'tsne':
self.compresser = TSNE(n_components=n_components, verbose=1)
elif self.method == 'lle':
self.compresser = LocallyLinearEmbedding(n_components=n_components,
n_jobs=4)
return self.compresser.fit_transform(X)
def LTSA(X, labels, imgs, n_neighbors, **kwargs):
# LTSA embedding of the dataset dataset
print("Computing LTSA embedding")
clf = LocallyLinearEmbedding(n_neighbors, n_components=2, method='ltsa')
t = time()
X_ltsa = clf.fit_transform(X)
print("Done. Reconstruction error: %g" % clf.reconstruction_error_)
plot_embedding(X_ltsa, labels, imgs,
"LTSA of the dataset (time %.2fs)" % (time() - t), **kwargs)
def lle(numComponents, neighbors=5, hessian=False):
# if there's time we can try changing n_neighbors
if hessian:
return LocallyLinearEmbedding(n_neighbors=neighbors,
n_components=numComponents,
method='hessian')
else:
return LocallyLinearEmbedding(n_neighbors=neighbors,
n_components=numComponents)
def draw_lle(matrix, spiral_density, layer_distance, k):
embedding = LocallyLinearEmbedding(n_components=2)
lle = embedding.fit_transform(matrix)
plt.clf()
plt.scatter(lle[:, 0], lle[:, 1])
title = "lle_spiral_density={0:.2f}_layer_distance={1:.2f}_k={2:.2f}.png".format(
spiral_density, layer_distance, k)
plt.title(title)
plt.savefig(title)
def __init__(self, method):
assert method in ['pca', 'tsne', 'lle']
self.method = method
if self.method == 'pca':
self.compresser = PCA(n_components=2)
elif self.method == 'tsne':
self.compresser = TSNE(n_components=2, verbose=1)
elif self.method == 'lle':
self.compresser = LocallyLinearEmbedding(n_components=2)
def data_transform(train, test):
pca = LocallyLinearEmbedding(n_components=80, n_neighbors=60)
train_tran = pca.fit_transform(train[:, :-1])
test_tran = pca.transform(test[:, :-1])
train_cat = np.hstack((train_tran, train[:, -1].reshape((-1, 1))))
test_cat = np.hstack((test_tran, test[:, -1].reshape((-1, 1))))
#print("explained variance ratio: %s" % str(pca.lambdas_))
pass
return train_cat, test_cat
def my_lle(X, y=None, l1=.1, n_components=2, **kwargs):
rrfs = RRFS(X.shape[1], hidden=n_components)
model = LocallyLinearEmbedding(n_components=n_components)
codes = model.fit_transform(X)
codes = (codes - np.min(codes)) / (np.max(codes) - np.min(codes))
#rrfs.train_representation_network(x_train, name=dataset+'_rep.hd5', epochs=1000)
score = rrfs.train_fs_network(X, rep=codes, l1=l1, epochs=300, loss='mse')
# sort the feature scores in an ascending order according to the feature scores
idx = np.argsort(score)[::-1]
return idx
def embed_lle(train, test, nn=10, method='standard'):
traintest = np.concatenate((train, test))
from sklearn.manifold import LocallyLinearEmbedding
lle = LocallyLinearEmbedding(n_neighbors=nn,
n_components=2,
method=method)
lle.fit(traintest)
X2d = lle.transform(traintest)
X2d = MinMaxScaler().fit_transform(X2d)
return X2d[:train.shape[0]], X2d[train.shape[0]:]
def runLLE_KMeans(self):
"""
Run sklearn-LocallyLinearEmbedding to reduce the dimensionality of the data
Cluster embedding with K-Means
"""
lle = LocallyLinearEmbedding(n_components=2)
self.dlle = lle.fit_transform(self.dataset)
self.kmeansLLE = KMeans(n_clusters=self.n_clusters,
random_state=0).fit_predict(self.dlle)
return self.dlle, self.kmeansLLE
def lle(data, d, k):
'''
input:data(ndarray):待降维数据
d(int):降维后数据维度
k(int):邻域内样本数
output:Z(ndarray):降维后数据
'''
lle = LocallyLinearEmbedding(n_components=d, n_neighbors=k)
Z = lle.fit_transform(data)
return Z
def nn_check(ppd):
for i in range(8, 26):
lle = LLE(n_components=3,
n_neighbors=i,
method='modified',
modified_tol=1e-12)
XT = lle.fit_transform(ppd)
print('running')
validity(XT, i)
print('done')
def __manifold_lle(pc, outcome, dim=2):
"""Fit Locally Linear Embedding.
:return: DataFrame of covariates"""
lle = LocallyLinearEmbedding(n_components=dim,
n_jobs=-1,
method='standard')
lle_out = lle.fit_transform(pc)
df_lle_out = pd.DataFrame(lle_out, columns=["D1", "D2"])
return df_lle_out
def pseudotimes_from_embedding(data_array, n_neighbors=None):
if n_neighbors is None:
n_neighbors = int(data_array.shape[0] * 0.5)
embedding = LocallyLinearEmbedding(n_components=1, n_neighbors=n_neighbors)
u, s, v = np.linalg.svd(data_array, full_matrices=1)
l = 2
denoised_data_array = np.dot(u[:, :l], np.dot(np.diag(s[:l]), v[:l, :]))
pseudotimes = embedding.fit_transform(denoised_data_array)
pseudotimes -= pseudotimes.min()
pseudotimes /= pseudotimes.max()
return pseudotimes
class SemiSupervisedGradientBoosting :
def __init__(self, max_depth=3, n_estimators=10, learning_rate=0.1,
min_samples_leaf=4, n_neighbors=5, n_components=2) :
self.GB = GradientBoosting.GradientBoosting(max_depth, n_estimators,
learning_rate, min_samples_leaf)
self.Transformator = LocallyLinearEmbedding(n_neighbors, n_components)
def fit_predict(self,Xl, y, Xu) :
print 'start collapse space'
delimeter = Xl.shape[0]
X_all = np.vstack((Xl, Xu))
X_all = self.Transformator.fit_transform(X_all)
X_l_t = X_all[:delimeter]
X_u_t = X_all[delimeter:]
del X_all
print 'start compute simalirity'
Sim = GradientBoosting.Simalirity(X_l_t, X_u_t)
print 'end compute simalirity'
del X_l_t, X_u_t
#Xl = X_all[:delimeter]
#Xu = X_all[delimeter:]
print 'end collapse space succesfully'
return self.GB.fit_predict(Xl, y, Xu, Sim)
def predict(self,X) :
return self.GB.predict(X)
def score (self, X, y) :
return self.GB.score(X, y)
def get_metastable_connections_from_gmm(data, gmm,
connection_estimation_method='max_path_distance_diff',
min_paths=3, distance='euclidean',
low_dimension_distances=True,
as_graph=False):
means = gmm.means_
memberships = gmm.predict(data)
if connection_estimation_method in ['max_path_distance_diff', 'connecting_paths', 'mst']:
if low_dimension_distances:
pca = PCA(n_components=2)
lle = LocallyLinearEmbedding(n_components=2,
n_neighbors=int(0.8*data.shape[0]))
distance_matrix = squareform(pdist(lle.fit_transform(data), distance))
else:
distance_matrix = squareform(pdist(data, distance))
weighted_graph = nx.Graph(distance_matrix)
else:
weighted_graph = None
return get_metastable_connections(data, means, memberships,
method=connection_estimation_method,
weighted_graph=weighted_graph, min_paths=3,
as_graph=as_graph)
from sklearn.manifold import LocallyLinearEmbedding
from astroML.datasets import fetch_sdss_specgals
from astroML.datasets import fetch_sdss_spectrum
data = fetch_sdss_specgals()
print data.dtype.names
ngals = 326
nwavel = 3855
plates = data['plate'][:ngals]
mjds = data['mjd'][:ngals]
fiberIDs = data['fiberID'][:ngals]
h_alpha = data['h_alpha_flux'][:ngals]
bptclass = data['bptclass'][:ngals]
specdata = np.zeros((ngals, nwavel))
i = 0
for plate, mjd, fiberID in zip(plates, mjds, fiberIDs):
tempdata = fetch_sdss_spectrum(plate, mjd, fiberID)
specdata[i, :] = tempdata.spectrum/tempdata.spectrum.mean()
i += 1
# Apply LLE
k = 7
for fignum, n in enumerate([2, 3]):
lle = LocallyLinearEmbedding(k, n)
lle.fit(specdata)
proj = lle.transform(specdata)
pl.subplot(2, 1, fignum+1)
pl.scatter(proj[:,0], proj[:,1], c=bptclass, s=50)
pl.colorbar()
pl.show()
def main():
parser = argparse.ArgumentParser(description=
'Perform Dimensionality Reduction')
parser.add_argument('--alg', type=str, default='MLLE',
help='Algorithm to reduce dimensionality.')
parser.add_argument('catalog', type=str,
help='Specify the catalog on which to perform DimReduce.')
args = parser.parse_args()
#dat = Table.read('catalogs/ZEST_catalog_colors.fits')
#training_sample = dat[0:10000]
#testing_sample = dat[10001:20000]
#zkeys = ['cc', 'aa', 'm20', 'gg']
base = os.path.basename(args.catalog)
filename = os.path.splitext(base)[0]
dat = Table.read(args.catalog)
mkeys = ['elipt', 'C', 'A_1a', 'G', 'M20']#
#dat.remove_column('color')
if 'color' not in dat.colnames:
if 'kaggle' in sample:
dat = prep_catalog.color_data2(dat, 'gz2class')
if 'direct' in sample:
dat = prep_catalog.color_data(dat, 'zclass')
dat.write(args.catalog, overwrite=True)
#dat = prep_catalog.adjust_asym(dat, mkeys[2])
#train, traincols, targets = prep_catalog.whiten_data(dat, mkeys)
n_neighbors = [10,12,15,20]
#n_neighbors = [7]
n_components = 3
for i, n_neigh in enumerate(n_neighbors):
if args.alg in ['MLLE', 'LLE', 'LTSA', 'HLLE']:
if args.alg == 'MLLE':
method = 'modified'
elif args.alg == 'LLE':
method = 'standard'
elif args.alg == 'LTSA':
method = 'ltsa'
elif args.alg == 'HLLE':
method = 'hessian'
#replace_panoptes(dat)
#pdb.set_trace()
#sample = 'directbig_panoptes'
X, y = prep_catalog.whiten_data(dat, mkeys)
(dat1, dat2),(thing1,thing2) = split_samples(dat, dat,[0.75, 0.35],
random_state=0)
(X_train, X_test), (y_train, y_test) = split_samples(X, y,
[0.75, 0.35], random_state=0)
y_train = simplify_classlabels(y_train)
y_test = simplify_classlabels(y_test)
#filename = 'modified_7_directbig_new'
X_train = X
y_train = simplify_classlabels(y)
#'''
#sample ='direct_zcut'
#Y_train, Y_test = open_previous_LLE(filename)
#cut = np.where(X1['REDSHIFT'] <= 0.05)
#X1_cut = X1[cut]
#QC_plots(X1_cut)
#Y_train = np.array(Y_train)[cut]
#col_train = np.array(col_train)[cut]
#X = Table(X)
#cut_out_mixedup_region(X, np.array(Y_train))
#'''
print "performing "+method+" LLE with",n_neigh,\
"nearest neighbors"
print "on training sample of",len(X_train),"objects"
t0 = time()
A = LLE(n_neigh, n_components, eigen_solver='auto', method=method)
error = A.fit(X_train).reconstruction_error_
Y_train = A.fit_transform(X_train)
Y_test = A.transform(X_train)
t1 = time()
#'''
metadata = {'method':method, 'N':n_neigh, 'd':n_components,
'error':error, 'time':t1-t0, 'sample':filename+'_total'}
save_dimreduce(dat, Y_train, y_train, metadata, filename+'_total')
#metadata = {'method':method, 'N':n_neigh, 'd':n_components,
# 'error':error, 'time':t1-t0, 'sample':filename+'_test'}
#save_dimreduce(X2, Y_test, y_test, metadata, filename+'_test')
# plot in 3D
plot_dimreduce_3D(Y_train, y_train[:,1], Y_test, y_test[:,1],
method, n_neigh, error, t1-t0, filename, two=False)
#====================================================================#
elif args.alg == 'ISO':
method='IsoMap'
print "performing IsoMap with",n_neigh,"nearest neighbors"
print "on training sample of",len(dat),"objects"
t0 = time()
A = Isomap(n_neigh, n_components, eigen_solver='dense')
error = A.fit(train).reconstruction_error()
Y = A.fit_transform(train)
#Y2 = A.transform(test)
t1 = time()
print "%s: %.2g sec" %(args.alg, t1-t0)
print "reconstruction error: ", error
print "begin plotting"
plot_dimreduce(Y, traincols, method, n_neigh, sample, axis=0)
plot_dimreduce(Y, traincols, method, n_neigh, sample, axis=1)
plot_dimreduce(Y, traincols, method, n_neigh, sample, axis=2)
plot_dimreduce_3D(Y, traincols, Y, traincols, method,
n_neigh, (t1-t0), error, sample)
elif args.alg == 'LDA':
print "performing LDA"
X, Xc, y = prep_catalog.whiten_data(dat, mkeys)
(X_train, X_test), (y_train, y_test) = split_samples(X, y,
[0.75, 0.25], random_state=0)
DRclf = LDA(3, priors=None)
#DRclf.fit(X_train, y_train)
DRtrain = DRclf.fit(X_train, y_train).transform(X_train)
DRtest = DRclf.fit(X_train, y_train).transform(X_test)
classes = np.unique(y_train)
colors = np.array(['darkred', 'red', 'lightsalmon',
'darkgreen', 'lightgreen', 'lightseagreen',
'indigo', 'darkviolet', 'plum'])
plot_LDA_3D(DRtrain, y_train, classes, colors, sample)
pdb.set_trace()
#classifiers = []
#predictions = []
#Nparams = np.arange(1, X.shape[1]+1)
#for nc in Nparams:
clf = LDA()
clf.fit(DRtrain, y_train)
y_pred = clf.predict(DRtest)
matchesLDA = (y_pred == y_test)
print np.sum(matchesLDA)
pdb.set_trace()
#------------------------------------------
from sklearn.neighbors import KNeighborsClassifier
knc = KNeighborsClassifier(5)
knc.fit(DRtrain, y_train)
y_pred = knc.predict(DRtest)
matchesKNN = (y_pred == y_test)
print np.sum(matchesKNN)
pdb.set_trace()
#------------------------------------------
from astroML.classification import GMMBayes
gmmb = GMMBayes(9)
gmmb.fit(DRtrain, y_train)
y_pred = gmmb.predict(DRtest)
matchesGMMB = (y_pred == y_test)
print np.sum(matchesGMMB)
pdb.set_trace()
#------------------------------------------
# plot the results
fig = plt.figure(figsize=(5, 2.5))
fig.subplots_adjust(bottom=0.15, top=0.95, hspace=0.0,
left=0.1, right=0.95, wspace=0.2)
# left plot: data and decision boundary
ax = fig.add_subplot(121)
pdb.set_trace()
im = ax.scatter(X[:, 3], X[:, 4], color=Xc, cmap=plt.cm.Spectral,
s=4, lw=0) #cmap=plt.cm.binary,, zorder=2
im.set_clim(-0.5, 1)
#im = ax.imshow(Z, origin='lower', aspect='auto',
# cmap=plt.cm.binary, zorder=1,
# extent=xlim + ylim)
#im.set_clim(0, 1.5)
#ax.contour(xx, yy, Z, [0.5], colors='k')
#ax.set_xlim(xlim)
#ax.set_ylim(ylim)
ax.set_xlabel('$G$')
ax.set_ylabel('$M20$')
#pred, true = classification_loss(predictions, y_test)
#completeness, contamination = completeness_contamination(pred, true)
pdb.set_trace()
#'''
#t0 = time()
#A = LDA(n_components, priors=None)
#Y = A.fit_transform(train, targets)
#Y2 = A.fit(train, targets).transform(train)
#t1 = time()
#print "%s: %.2g sec" %(args.alg, t1-t0)
predict = A.predict(train)
#print "Predicted classes:", predict
#pdb.set_trace()
#pdb.set_trace()
#'''
plot_LDA_3D(Y2, targets, classes, colors, sample)
plot_LDA(Y2, targets, classes, colors, sample, axis=0)
plot_LDA(Y2, targets, classes, colors, sample, axis=1)
plot_LDA(Y2, targets, classes, colors, sample, axis=2)
pdb.set_trace()
dic_cl[st_name][5] += int(fields[6].lower().strip().strip('"'))
dic_cl[st_name][6] += int(fields[7].lower().strip().strip('"'))
f.close()
import numpy as np
N = len(dic_cl.items())
X = np.zeros((N, 7))
for i, (key, val) in enumerate(dic_cl.iteritems()):
X[i, :] = dic_cl[key]
from sklearn.manifold import LocallyLinearEmbedding
from sklearn.preprocessing import scale
lle = LocallyLinearEmbedding(n_components=3, n_neighbors=20)
print X.max(axis=0)
Y3 = lle.fit_transform(scale(X))
Y3 -= Y3.min(axis=0)
print len(dic_cl.items())
lle = LocallyLinearEmbedding(n_components=1, n_neighbors=20)
Y1 = lle.fit_transform(X)
Y1 -= Y1.min()
o1 = open("1-d.csv", "w")
o3 = open("3-d.csv", "w")
for i, (key, val) in enumerate(dic_cl.iteritems()):
o1.write("%s,%f\n" % (key, Y1[i - 1]))
o3.write("%s,%s\n" % (key, ",".join(map(str, Y3[i - 1, :]))))
o1.close()
features_train_transformed = selector.transform(features_train_transformed).toarray()
return features_train_transformed, lables, vectorizer, selector, le, features
# nFeatures = np.arange(50, 1000, 50)
nLocally_Linear = np.arange(20, 200, 20)
data = {}
for k in nLocally_Linear:
features, labels, vectorizer, selector, le, features_data = preprocess("pkl/article_2_people.pkl", "pkl/lable_2_people.pkl")
features_train, features_test, labels_train, labels_test = cross_validation.train_test_split(features, labels, test_size=0.1, random_state=42)
t0 = time()
ll = LocallyLinearEmbedding(n_neighbors=15, n_components=k, eigen_solver='auto')
ll.fit(features_train)
print ("Dimension Reduction time:", round(time()-t0, 3), "s")
features_train = ll.transform(features_train)
features_test = ll.transform(features_test)
for name, clf in [
('AdaBoostClassifier', AdaBoostClassifier(algorithm='SAMME.R')),
('BernoulliNB', BernoulliNB(alpha=1)),
('GaussianNB', GaussianNB()),
('DecisionTreeClassifier', DecisionTreeClassifier(min_samples_split=100)),
('KNeighborsClassifier', KNeighborsClassifier(n_neighbors=50, algorithm='ball_tree')),
('RandomForestClassifier', RandomForestClassifier(min_samples_split=100)),
('SVC', SVC(kernel='linear', C=1))
iso = Isomap(n_components=3, n_neighbors=15) fdata = iso.fit_transform(digits["data"]) fig = plt.figure() ax = fig.add_subplot(111, projection="3d") plt.scatter(fdata[:, 0], fdata[:, 1], zs=fdata[:, 2], c=digits["target"], s=100) plt.show() # LLE from sklearn.manifold import LocallyLinearEmbedding lle = LocallyLinearEmbedding(n_neighbors=15, n_components=3, method="modified") fig = plt.figure() fdata = lle.fit_transform(digits["data"]) ax = fig.add_subplot(111, projection="3d") plt.scatter(fdata[:, 0], fdata[:, 1], zs=fdata[:, 2], c=digits["target"], s=100) plt.show() # MDS from sklearn.manifold import MDS mds = MDS(n_components=3) fig = plt.figure() fdata = mds.fit_transform(digits["data"])
from sklearn.manifold import LocallyLinearEmbedding
import numpy as np
import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt
tetra_freq = np.load('tetrafreq.npy')
phylum_index = np.load('phylumIndex.npy')
phylum_names = np.load('phylumNames.npy')
lle = LocallyLinearEmbedding(n_components=2)
lle_result = lle.fit_transform(tetra_freq)
plt.figure()
for c, i, name in zip ("bgrcmykw", list(range(7, -1, -1)), phylum_names):
plt.scatter(lle_result[phylum_index == i, 0], lle_result[phylum_index == i, 1], c=c, label=name)
plt.title('LLE of tetranucleotide')
plt.legend(loc=3, fontsize=10)
plt.savefig('LLE.png')
reducedImages = pca.fit_transform(trimmedImages)
elif sys.argv[1] == '-isomap':
trimmedImages = []
for i in range(len(images)):
images[i] = np.reshape(images[i], (-1))
images[i] = images[i][:minSize]
trimmedImages.append(images[i])
isomap = Isomap(n_components=136)
reducedImages = isomap.fit_transform(trimmedImages)
elif sys.argv[1] == '-lle':
trimmedImages = []
for i in range(len(images)):
images[i] = np.reshape(images[i], (-1))
images[i] = images[i][:minSize]
trimmedImages.append(images[i])
lle = LocallyLinearEmbedding(n_components=136)
reducedImages = lle.fit_transform(trimmedImages)
# Do cross-fold validation
kf = KFold(len(images), n_folds=2)
minAreas = {}
maxAreas = {}
avgAreas = {}
totals = {}
for train_index, test_index in kf:
xTrain = reducedImages[train_index]
yTrain = labels[train_index]
clf = OneVsRestClassifier(LinearSVC(), 4)
clf.fit(xTrain, yTrain)
xTest = reducedImages[test_index]
yTest = labels[test_index]
def localLinearEmbedding(X, y): lle = LocallyLinearEmbedding(n_components = 1, eigen_solver = "dense") lle.fit(X) transformX = lle.transform(X) return transformX
def plot2d(X, y, scale=True, normalize=False, embedding='pca', title=''):
"""
Plot data transformed into two dimensions by PCA.
PCA transforms into a new embedding dimension such that
the first dimension contains the maximal variance and following
dimensions maximal remaining variance.
This shoudl spread the observed n-dimensional data maximal. This
is unsupervised and will not consider target values.
"""
if (scale):
scaler = StandardScaler()
X = scaler.fit_transform(X)
if (normalize):
normalizer = Normalizer(norm='l2')
X = normalizer.fit_transform(X)
if (embedding is 'pca'):
pca = PCA(n_components=2)
X_transformed = pca.fit_transform(X)
elif (embedding is 'isomap'):
isomap = Isomap(n_components=2, n_neighbors=20)
X_transformed = isomap.fit_transform(X)
elif (embedding is 'lle' ):
lle = LocallyLinearEmbedding(n_components=2, n_neighbors=5)
X_transformed = lle.fit_transform(X)
elif (embedding is 'tsne'):
t_sne = TSNE(n_components=2)
X_transformed = t_sne.fit_transform(X)
elif (embedding is 'spectral'):
se = SpectralEmbedding(n_components=2)
X_transformed = se.fit_transform(X)
elif (embedding is 'mds'):
mds = MDS(n_components=2)
X_transformed = mds.fit_transform(X)
elif (embedding is 'gallery'):
plt.figure(1)
plt.subplot(231)
plt.title('pca')
X_t = PCA(n_components=2).fit_transform(X)
plt.scatter(X_t[:,0 ], X_t[:, 1], c=y)
plt.subplot(232)
plt.title('isomap')
X_t = Isomap(n_neighbors=20).fit_transform(X)
plt.scatter(X_t[:,0 ], X_t[:, 1], c=y)
plt.subplot(233)
plt.title('lle')
X_t = LocallyLinearEmbedding(n_neighbors=20).fit_transform(X)
plt.scatter(X_t[:,0 ], X_t[:, 1], c=y)
plt.subplot(234)
plt.title('tsne')
X_t = TSNE().fit_transform(X)
plt.scatter(X_t[:,0 ], X_t[:, 1], c=y)
plt.subplot(235)
plt.title('spectral')
X_t = SpectralEmbedding().fit_transform(X)
plt.scatter(X_t[:,0 ], X_t[:, 1], c=y)
plt.subplot(236)
plt.title('mds')
X_t = MDS().fit_transform(X)
plt.scatter(X_t[:,0 ], X_t[:, 1], c=y)
plt.suptitle('Gallery transforms ' + title)
return plt
else:
raise ValueError("Choose between pca, isomap and tsne")
plt.title(title + ' ' + embedding + ' plot')
sc = plt.scatter(X_transformed[:, 0], X_transformed[:, 1], c=y)
plt.colorbar(sc)
return plt
n_samples, n_features = D.shape
n_neighbors = 10
#----------------------------------------------------------------------
# Isomap projection
print "Computing Isomap embedding"
t0 = time()
D_iso = Isomap(n_neighbors, n_components=2).fit_transform(D_scaled)
print "Done in time %.2fs " % (time() - t0)
#----------------------------------------------------------------------
# Locally linear embedding
n_neighbors = 35
print "Computing LLE embedding"
clf = LocallyLinearEmbedding(n_neighbors, n_components=2,
method='modified')
t0 = time()
D_lle = clf.fit_transform(D_scaled)
print "Done in time %.2fs " % (time() - t0)
print "Reconstruction error: %g" % clf.reconstruction_error_
#----------------------------------------------------------------------
# kernel PCA
print "Computing kPCA embedding"
kpca = KernelPCA(n_components=2, kernel="rbf", gamma=0.0028942661247167516)
t0 = time()
D_kpca = kpca.fit_transform(D_scaled)
print "Done in time %.2fs " % (time() - t0)
plot_embedding(D_pca, 1, rescale=None, title="PCA projection")
plot_embedding(D_iso, 2, rescale=None, title="Isomap projection")
from __future__ import division
import sys
from sklearn.decomposition import PCA
from sklearn.manifold import Isomap
from sklearn.manifold import LocallyLinearEmbedding
from sklearn import preprocessing
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.cm as cm
from mpl_toolkits.mplot3d import Axes3D
import random
from colorsys import hsv_to_rgb
pca = PCA(n_components=2)
isomap = Isomap(n_components=2)
lle = LocallyLinearEmbedding(n_components=2)
data = np.genfromtxt('data01_small.txt', delimiter=',')
pca_xform = pca.fit_transform(data)
isomap_xform = isomap.fit_transform(data)
lle_xform = lle.fit_transform(data)
label = [0]*100+[1]*100
rgbs = [(0.5,0,0), (0,0.5,0)]
plt.figure()
xs = pca_xform[:,0]
ys = pca_xform[:,1]
ax = plt.subplot(111)
for i in xrange(len(xs)):
ax.text(xs[i], ys[i], str(label[i]), color=rgbs[label[i]], fontdict={'weight': 'bold', 'size': 9})
t = (max(xs)-min(xs))*0.1
def lle(X=None, W=None, num_vecs=None, k=None):
embedder = LocallyLinearEmbedding(n_neighbors=k, n_components=num_vecs)
return embedder.fit_transform(X)
# Build the output arrays
cells = opts.high / opts.step
lle_gmm_results = np.zeros((cells,opts.iters))
D = scale(X)
n_samples, n_features = D.shape
# chosen by hyperparam search in a separate test.
n_neighbors = 35
# For the specified number of principal components, do the clustering
dimension_list = range(opts.low, opts.high + 1, opts.step)
data_files = []
for i in dimension_list:
index = (i / opts.step) - 1
lle = LocallyLinearEmbedding(n_neighbors, n_components=i, method='standard')
X_lle = lle.fit_transform(D)
for j in range(0,opts.iters,1):
gaussmix = GMM(n_components=true_k, covariance_type='tied', n_init=10, n_iter=1000)
gaussmix.fit(X_lle)
gaussmix_labels = gaussmix.predict(X_lle)
homog = metrics.homogeneity_score(labels[:,0], gaussmix_labels)
print "Gaussian mixture homogeneity: %0.3f" % homog
test_result = {"Model": "LLE", "Dimension": i, "Homogeneity": homog}
index = pd.Index([0], name='rows')
data_files.append(pd.DataFrame(data=test_result,index=index))
# Save the data to a file:
print "...Done"
'''
train PCA basis based on training.txt and output dimension-reduced coefficients for both training.txt and testing.txt
'''
from __future__ import division
import sys
from sklearn.decomposition import PCA
from sklearn.manifold import Isomap
from sklearn.manifold import LocallyLinearEmbedding
from sklearn import preprocessing
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.cm as cm
from mpl_toolkits.mplot3d import Axes3D
import random
from colorsys import hsv_to_rgb
final_dim = 30
data = np.genfromtxt("100examples.txt", delimiter=',')
pca = PCA(n_components=final_dim)
isomap = Isomap(n_components=final_dim)
lle = LocallyLinearEmbedding(n_components=final_dim)
data_xformed = lle.fit_transform(data)
np.savetxt("lle_data_30_dims.txt", data_xformed, delimiter=',')
#03-02.py
X, y = preprocess(data, shuffle=False, n_samples=1000, normalization=None)
from sklearn.manifold import LocallyLinearEmbedding
lle = LocallyLinearEmbedding(n_neighbors=15,
n_components=3, method='modified')
X_proj = lle.fit_transform(X)
three_component_plot(X_proj[:, 0], X_proj[:, 1], X_proj[:, 2], y, labels, trim_outliers=True)
def __init__(self, max_depth=3, n_estimators=10, learning_rate=0.1,
min_samples_leaf=4, n_neighbors=5, n_components=2) :
self.GB = GradientBoosting.GradientBoosting(max_depth, n_estimators,
learning_rate, min_samples_leaf)
self.Transformator = LocallyLinearEmbedding(n_neighbors, n_components)