def compute_silhouette_score(clusters):
"""
Compute the euclidean silhouette score and the cosine silhouette score. Return the scores.
:param clusters: clusters assignment for each tweet
:type clusters: list
:return: the silhouette scores
:rtype: tuple
"""
# Load the files
tfidf_matrix = pickle.load(open('TF-IDF Matrix - ' + str(n_data) + ' Tweets.p', 'rb'))
# Compute the Silhouette Score
start = timer()
distance = 1 - cosine_similarity(tfidf_matrix)
euclidean_silhouette_score = silhouette_score(tfidf_matrix, clusters, metric='euclidean')
cosine_silhouette_score = silhouette_score(distance, clusters, metric='precomputed')
end = timer()
print('Silhouette Score (Euclidean): %.4f' % euclidean_silhouette_score)
print('Silhouette Score (Cosine): %.4f' % cosine_silhouette_score)
print('Obtained the Silhouette Score in %.2f seconds' % (end - start))
txt_file.write('Silhouette Score (Euclidean): %.4f. \n' % euclidean_silhouette_score)
txt_file.write('Silhouette Score (Cosine): %.4f. \n' % cosine_silhouette_score)
txt_file.write('Obtained the Silhouette Score in %.2f seconds. \n' % (end - start))
return euclidean_silhouette_score, cosine_silhouette_score
def test_sihouette(self):
n1 = np.array([[1,2,1], [1,3,1], [7,8,2], [7,9,2], [13,19,3]])
print(Silhouette.score(n1))
print(silhouette_score(n1, n1[:,-1]))
n2 = np.array([[1,2,1], [1,3,2], [7,8,2], [7,9,1], [13,19,3]])
print(Silhouette.score(n2))
print(silhouette_score(n2, n2[:,-1]))
def get_constant_height_labels(clustering, n_clusters=None):
"""
use silhouette analysis to select the best heigh to cut a linkage matrix
:df: a correlation matrix
parse_heatmap: int (optional). If defined, devides the columns of the
heatmap based on cutting the dendrogram
"""
N_variables = len(clustering['reorder_vec'])
scores = []
if n_clusters is None:
for k_clusters in range(2,N_variables//3):
labels = cut_tree(clustering['linkage'], n_clusters=k_clusters)
try:
score = silhouette_score(clustering['distance_df'],
labels.ravel(), metric='precomputed')
except ValueError:
continue
scores.append((k_clusters,score))
best_k = max(scores, key=lambda x: x[1])[0]
labels = cut_tree(clustering['linkage'], n_clusters=best_k)
else:
labels = cut_tree(clustering['linkage'], n_clusters=n_clusters)
score = silhouette_score(clustering['distance_df'],
labels, metric='precomputed')
scores.append((n_clusters, score))
labels = reorder_labels(labels.flatten(), clustering['linkage'])
# comparison
MI = adjusted_mutual_info_score(labels, clustering['labels'])
return labels, scores, MI
def bench_k_means(estimator, data, labels):
t0 = time()
estimator.fit(data)
print("time to fit: {:.5}".format(time() - t0))
homogenity = metrics.homogeneity_score(labels, estimator.labels_)
completeness = metrics.completeness_score(labels, estimator.labels_)
v_measure = metrics.v_measure_score(labels, estimator.labels_)
print("homogenity {:.5}, completeness {:.5}, v_measure_score {:.5}".format(
homogenity, completeness, v_measure)
)
adj_rand_score = metrics.adjusted_rand_score(
labels, estimator.labels_
)
print("adjusted_rand_score {:.5}".format(adj_rand_score))
adj_mutual_info_score = metrics.adjusted_mutual_info_score(
labels, estimator.labels_
)
print("adjusted_mutual_info_score {:.5}".format(
adj_mutual_info_score)
)
silhouette_score = metrics.silhouette_score(
data, estimator.labels_, metric='euclidean'
)
print("silhouette_score {:.5}".format(
metrics.silhouette_score(data, estimator.labels_,
metric='euclidean'))
)
return [
homogenity, completeness, v_measure, adj_rand_score,
adj_mutual_info_score, silhouette_score
]
def clustering_drawing():
X,Tag = getData()
n = 3
kmeans_model = KMeans(n_clusters = n).fit(X)
labels = kmeans_model.labels_
score = metrics.silhouette_score(X, labels, metric='euclidean')
scoreList = [score]
nList = [3,4,5,6,7,8,9]
for i in range(4,10):# 聚类4-10类循环
# print i
kmeans_model_temp = KMeans(n_clusters=i).fit(X)
labels_temp = kmeans_model_temp.labels_
score_temp = metrics.silhouette_score(X, labels_temp, metric='euclidean')
print i,score_temp
scoreList.append(float(score_temp))
if float(score_temp) > score:
kmeans_model = kmeans_model_temp
labels = labels_temp
n = i
print n,labels
plt.axis([3,9,0.8,1.0])
plt.plot(nList, scoreList, 'r--')
plt.show()
def print_cluster(clusterTrainClass, labels, clusterTestStory):
print("Homogeneity: %0.3f" % metrics.homogeneity_score(clusterTrainClass, labels))
print("Completeness: %0.3f" % metrics.completeness_score(clusterTrainClass, labels))
print("V-measure: %0.3f" % metrics.v_measure_score(clusterTrainClass, labels))
print("Adjusted Rand Index: %0.3f" % metrics.adjusted_rand_score(clusterTrainClass, labels))
print("Adjusted Mutual Information: %0.3f" % metrics.adjusted_mutual_info_score(clusterTrainClass, labels))
print "Silhouette Coefficient:"
print metrics.silhouette_score(clusterTestStory, labels, metric='euclidean')
def benchmark(estimator, name, data):
t0 = time()
estimator.fit(data)
print('% 9s %.2fs %i %.3f'
% (name, (time() - t0), estimator.inertia_,
metrics.silhouette_score(data, estimator.labels_,
metric='euclidean',
sample_size=None)))
return [time() - t0, estimator.inertia_, metrics.silhouette_score(data, estimator.labels_, metric='euclidean', sample_size=None)]
def drawwDBSCAN(newarray,comparearray,cityname):
X = StandardScaler().fit_transform(newarray)
# print newarray
# print "#########"
# print X
# X = newarray
##############################################################################
# Compute DBSCAN
db = DBSCAN(eps=0.3, min_samples=10).fit(X)
core_samples_mask = np.zeros_like(db.labels_, dtype=bool)
core_samples_mask[db.core_sample_indices_] = True
labels = db.labels_
# Number of clusters in labels, ignoring noise if present.
n_clusters_ = len(set(labels)) - (1 if -1 in labels else 0)
print('Estimated number of clusters: %d' % n_clusters_)
print("Silhouette Coefficient: %0.3f"
% metrics.silhouette_score(X, labels))
##############################################################################
# Plot result
matplotlib.style.use('ggplot')
# Black removed and is used for noise instead.
unique_labels = set(labels)
colors = plt.cm.Spectral(np.linspace(0, 1, len(unique_labels)))
for k, col in zip(unique_labels, colors):
if k == -1:
# Black used for noise.
col = 'k'
class_member_mask = (labels == k)
xy = X[class_member_mask & core_samples_mask]
plt.plot(xy[:, 0], xy[:, 1], 'o', markerfacecolor=col,
markeredgecolor='k', markersize=14)
xy = X[class_member_mask & ~core_samples_mask]
plt.plot(xy[:, 0], xy[:, 1], 'o', markerfacecolor=col,
markeredgecolor='k', markersize=6)
plt.title('Estimated number of clusters: %d' % n_clusters_)
imgname = "./clusterimage/hourcondimention/" +cityname+'.png'
fig = plt.gcf()
fig.set_size_inches(16.5, 12.5)
fig.savefig(imgname)
ScandARI = drawlableCLuster(comparearray,labels,cityname.split('_')[3])
print ScandARI
with open('summary_hour_total_dimention.csv','a') as f:
write = csv.writer(f)
# write.writerow(['name','clusters','SC'])
write.writerow([cityname,n_clusters_,metrics.silhouette_score(X, labels, metric='sqeuclidean')])
write.writerow(["hour_dimention_twitterinfo"+cityname.split('_')[3],ScandARI[0],ScandARI[1],ScandARI[2]])
def eval_perf(self):
X_tst, y_tst = self.data.get_test_set()
code = self.dbn.f_code(X_tst)
from sklearn import metrics
sil_c = metrics.silhouette_score(code, y_tst)
sil_X = metrics.silhouette_score(X_tst, y_tst)
print 'Silhouette code y', sil_c
print 'Silhouette X y', sil_X
def cluster_kmeans1():
dataset = datasets.load_iris()
X = dataset.data
kmeans_model = KMeans(n_clusters=4,random_state=1).fit(X)
labels = kmeans_model.labels_
print X
print kmeans_model.cluster_centers_
print labels
print metrics.silhouette_score(X,labels,metric="euclidean")
def kmeans(input_file, n_clusters, Output):
lvltrace.lvltrace("LVLEntree dans kmeans unsupervised")
ncol=tools.file_col_coma(input_file)
data = np.loadtxt(input_file, delimiter=',', usecols=range(ncol-1))
X = data[:,1:]
y = data[:,0]
sample_size, n_features = X.shape
k_means=cluster.KMeans(init='k-means++', n_clusters=n_clusters, n_init=10)
k_means.fit(X)
reduced_data = k_means.transform(X)
values = k_means.cluster_centers_.squeeze()
labels = k_means.labels_
k_means_cluster_centers = k_means.cluster_centers_
print "#########################################################################################################\n"
#print y
#print labels
print "K-MEANS\n"
print('homogeneity_score: %f'%metrics.homogeneity_score(y, labels))
print('completeness_score: %f'%metrics.completeness_score(y, labels))
print('v_measure_score: %f'%metrics.v_measure_score(y, labels))
print('adjusted_rand_score: %f'%metrics.adjusted_rand_score(y, labels))
print('adjusted_mutual_info_score: %f'%metrics.adjusted_mutual_info_score(y, labels))
print('silhouette_score: %f'%metrics.silhouette_score(X, labels, metric='euclidean', sample_size=sample_size))
print('\n')
print "#########################################################################################################\n"
results = Output+"kmeans_scores.txt"
file = open(results, "w")
file.write("K-Means Scores\n")
file.write("Homogeneity Score: %f\n"%metrics.homogeneity_score(y, labels))
file.write("Completeness Score: %f\n"%metrics.completeness_score(y, labels))
file.write("V-Measure: %f\n"%metrics.v_measure_score(y, labels))
file.write("The adjusted Rand index: %f\n"%metrics.adjusted_rand_score(y, labels))
file.write("Adjusted Mutual Information: %f\n"%metrics.adjusted_mutual_info_score(y, labels))
file.write("Silhouette Score: %f\n"%metrics.silhouette_score(X, labels, metric='euclidean', sample_size=sample_size))
file.write("\n")
file.write("True Value, Cluster numbers, Iteration\n")
for n in xrange(len(y)):
file.write("%f, %f, %i\n"%(y[n],labels[n],(n+1)))
file.close()
import pylab as pl
from itertools import cycle
# plot the results along with the labels
k_means_cluster_centers = k_means.cluster_centers_
fig, ax = plt.subplots()
im=ax.scatter(X[:, 0], X[:, 1], c=labels, marker='.')
for k in xrange(n_clusters):
my_members = labels == k
cluster_center = k_means_cluster_centers[k]
ax.plot(cluster_center[0], cluster_center[1], 'w', color='b',
marker='x', markersize=6)
fig.colorbar(im)
plt.title("Number of clusters: %i"%n_clusters)
save = Output + "kmeans.png"
plt.savefig(save)
lvltrace.lvltrace("LVLsortie dans kmeans unsupervised")
def acc_silhouette(X, lbls_true, lbls_pred, reject, strat_lbl_inds, use_strat=False, metric='euclidean'):
if use_strat:
dists = sc.distances(X[:, strat_lbl_inds], gene_ids=np.arange(strat_lbl_inds.size), metric=metric )
sil = metrics.silhouette_score(dists, lbls_pred[strat_lbl_inds], metric='precomputed')
perc = np.int(np.float(len(strat_lbl_inds))/np.float(lbls_true.size) * 100.0)
desc = ('Silhouette (strat={0},{1})'.format(perc, metric), 'Silhouette ({0})'.format(metric))
else:
dists = sc.distances(X, gene_ids=np.arange(X.shape[1]), metric=metric )
sil = metrics.silhouette_score(dists, lbls_pred, metric='precomputed')
desc = ('Silhouette ({0})'.format(metric), 'Silhouette ({0})'.format(metric))
return sil, desc
def fit(self, X, Y=None):
proj = skl_cluster.KMeans(**self.params)
if isinstance(X, Table):
proj = proj.fit(X.X, Y)
proj.silhouette = silhouette_score(X.X, proj.labels_)
else:
proj = proj.fit(X, Y)
proj.silhouette = silhouette_score(X, proj.labels_)
proj.inertia = proj.inertia_ / len(X)
cluster_dist = Euclidean(proj.cluster_centers_)
proj.inter_cluster = np.mean(cluster_dist[np.triu_indices_from(cluster_dist, 1)])
return KMeansModel(proj, self.preprocessors)
def bench_k_means(estimator, name, data, silhouette_results):
t0 = time()
estimator.fit(data)
print('init\t\ttime\tinertia\thomo\tcompl\tv-meas\tARI\tAMI\tsilhouette')
print('% 9s\t %.2fs\t%i\t%.3f\t%.3f\t%.3f\t%.3f\t%.3f\t%.3f' % \
(name, (time() - t0), \
estimator.inertia_, \
metrics.homogeneity_score(labels, estimator.labels_), \
metrics.completeness_score(labels, estimator.labels_), \
metrics.v_measure_score(labels, estimator.labels_), \
metrics.adjusted_rand_score(labels, estimator.labels_), \
metrics.adjusted_mutual_info_score(labels, estimator.labels_), \
metrics.silhouette_score(data, estimator.labels_, metric='euclidean')))
return str(metrics.silhouette_score(data,estimator.labels_, metric='euclidean'))
def cluster_driver(a_driver):
# print a_driver['DStats']
# print "#############################DStats Above#################################ValueError: zero-size array to reduction operation minimum which has no identity#################"
X = StandardScaler().fit_transform(a_driver['DStats'])
# print X
# print "DStats are.....::" , a_driver['DStats']
# print "X is...........::" ,['AvgDistDel', 'AvgACosDel', 'SDevDistDel', 'SDevACosDel','TotalTime','SkewDistDel','SkewACosDel'] X
# print "############################Scaled X Above###################################################"
# db = KMeans(n_clusters=20,n_jobs = -1).fit(X)
db = DBSCAN(eps=0.45).fit(X)
# core_samples_mask = np.zeros_like(db.labels_, dtype=bool)
# core_samples_mask[db.core_sample_indices_] = True
labels = db.labels_
# n_clusters_ = len(set(labels)) - (1 if -1 in labels else 0)
print "###############################################################################"
# print('Estimated number of clusters: %d' % n_clusters_)
# print 'Count of Predicts::', len(X)
print("Silhouette Coefficient: %0.3f" % metrics.silhouette_score(X, labels,metric="mahalanobis"))
# print "##############################DBSCAN X Below#################################################"
# print X G:/Continuing Education/Research & Presentations/Self - Machine Learning/Kaggle/DriverTelemetricAnalysis-AXA/'
# try:
return (metrics.silhouette_samples(X, labels,metric="mahalanobis")+1)/2
def save_cluster_metrics(self, points, predictions, key, level2_mode = False): try: silhoutte_global = metrics.silhouette_score(points, predictions, metric='euclidean') silhoutte_weighted = utils.silhoutte_weighted(points, predictions) self.silhouette_scores_global[key] = silhoutte_global self.silhouette_scores_weighted[key] = silhoutte_weighted if level2_mode: self.level2_silhoutte_global.append(silhoutte_global) self.level2_silhoutte_weighted.append(silhoutte_weighted) except ValueError as e: pass # dunn_scores = cluster_evaluation.dunn_index(points, predictions, means) dunn_scores = [0, 0, 0] if (dunn_scores[0] is not None) and (dunn_scores[1] is not None) and (dunn_scores[2] is not None): self.dunn_scores_1[key] = dunn_scores[0] self.dunn_scores_2[key] = dunn_scores[1] self.dunn_scores_3[key] = dunn_scores[2] if level2_mode: self.level2_dunn_1.append(dunn_scores[0]) self.level2_dunn_2.append(dunn_scores[1]) self.level2_dunn_3.append(dunn_scores[2])
def find_best(df, cls, norm):
'''
INPUTS: Pandas DataFrame, String of which dataset is being used,
Boolean if data is normalized
OUTPUTS: Prints score to screen. Saves model to RESULTS_DIR
'''
if norm == True:
df = StandardScaler(copy=False).fit_transform(df)
files = [f for f in os.listdir(RESULTS_DIR)
if f.endswith('{}_{}.pkl'.format(cls, norm)) and
if f.startswith('k')]]
scores = []
for f in files:
model = pickle.load(open(RESULTS_DIR + f, 'rb'))
labels = model.predict(df)
score = silhouette_score(df.values, labels, sample_size=10000)
name = f.split('_')[1]
scores.append((score, name))
print "{} {} {} {}".format(f.split('_')[1], float(score), cls, norm)
del labels
del model
ranked_scores = sorted(scores, reverse=True)
ranked_scores = [(item[1], item[0]) for item in ranked_scores]
with open('{}_{}.pkl'.format(cls, norm), 'wb') as f:
pickle.dump(ranked_scores, f)
for item in ranked_scores: print item
def Create_Ext_Agg_cluster(self,stem,stop,processing,remS):
Allrow_dicts=data_pkg.FileHandling.read_csv(self.ExtStringCSv)
Allstrings=list()
#Allstrings=[rowdict_str["Text_original"] for rowdict_str in Allrow_dicts]
for row_dict in Allrow_dicts:
if self.POS =="ALL_EXT":
Stringrow=row_dict["Text_original"]+row_dict["Adj_Extended"]+row_dict["Noun_Extended"] +row_dict["Verb_Extended"]
Allstrings.append(Stringrow)
else:
Stringrow=row_dict["Adj"]+row_dict["Adj_Extended"]+row_dict["Noun"]+row_dict["Noun_Extended"]#+row_dict["Verb"]#+row_dict["Verb_Extended"]
Allstrings.append(Stringrow)
Allstrings_process=[preprocess_text.preprocess(string_text, stem,stop) for string_text in Allstrings]
if remS:
Allstrings_process=[preprocess_text.removeS(text) for text in Allstrings_process]
vectorizer = CountVectorizer()
term_doc=vectorizer.fit_transform(Allstrings_process)
#-------------------------- feature_names=vectorizer.get_feature_names()
#--z---------------------------------------------- Array=term_doc.toarray
if self.affinity=='euclidean':
Agg_cluster=AgglomerativeClustering(n_clusters=self.num_cluster,affinity='euclidean')
if self.affinity=='cosine':
Agg_cluster=AgglomerativeClustering(n_clusters=self.num_cluster,linkage='average',affinity='cosine')
Res_Labels=Agg_cluster.fit_predict(term_doc.toarray())
self.cluster_tup_list=self.tuple_Ext_cluster_doc(Res_Labels,Allstrings,Allrow_dicts)
#term_doc_lsa = lsa.fit_transform(term_doc)
print type (term_doc)
self.metric=metrics.silhouette_score(term_doc.toarray(), Res_Labels, metric=self.affinity)
print Res_Labels
print("n_samples: %d, n_features: %d" % term_doc.shape)
def ap(data):
X = data
af = AffinityPropagation(
damping=0.8,
max_iter=200,
convergence_iter=15,
preference=None,
affinity='euclidean',
verbose=True).fit(X)
cluster_centers_indices = af.cluster_centers_indices_
labels = af.labels_
n_clusters_ = len(cluster_centers_indices)
print('Estimated number of clusters: %d' % n_clusters_)
# print(
# "Homogeneity: %0.3f" % metrics.homogeneity_score(labels_true, labels))
# print("Completeness: %0.3f" % metrics.completeness_score(
# labels_true, labels))
# print("V-measure: %0.3f" % metrics.v_measure_score(labels_true, labels))
# print("Adjusted Rand Index: %0.3f" % metrics.adjusted_rand_score(
# labels_true, labels))
# print("Adjusted Mutual Information: %0.3f" %
# metrics.adjusted_mutual_info_score(labels_true, labels))
print("Silhouette Coefficient: %0.3f" % metrics.silhouette_score(
X, labels, metric='sqeuclidean'))
def __initialize_clusters(self):
# Load the data
self.companies, self.descriptions, self.company_idx_map = load_data(self.data_dir, has_header=True)
# Vectorize the data using TF-IDF to help reduce dimensionality
self.vectorizer = TfidfVectorizer(max_df=0.5, stop_words='english', use_idf=True, ngram_range=(1, 2)) #min_df=2
X = self.vectorizer.fit_transform(self.descriptions)
self.instance_vector_array = X.toarray()
print("n_samples: %d, n_features: %d" % X.shape)
# Initialize K-means algorithm preferences
print("Initializing clusters, this takes a few seconds ...")
self.km = KMeans(n_clusters=self.k, init='k-means++', max_iter=100, n_init=1, verbose=False)
# Use k-means to generate clusters
self.km.fit(X)
# initialize results dictionary
labels = self.km.labels_
for i in range(0, self.k, 1):
self.results[i] = []
# assign results by label
for i in range(0, len(labels), 1):
self.results[labels[i]].append(self.companies[i])
self.company_to_cluster[self.companies[i].name] = labels[i]
print("Silhouette Coefficient: %0.3f"
% metrics.silhouette_score(X, self.km.labels_, sample_size=1000))
# Write cluster results to Output file
output_results(self.km, self.k, self.vectorizer, self.results)
pass
def cluster_evaluation(D, y_true, n_clusters, eps=0.8, min_samples=10):
##############################################################################
# Extract Y true
labels_true = y_true
##############################################################################
# transform distance matrix into a similarity matrix
S = 1 - D
##############################################################################
# compute DBSCAN
#db = DBSCAN(eps=eps, min_samples=min_samples).fit(S)
db = Ward(n_clusters=n_clusters).fit(S)
#core_samples = db.core_sample_indices_
labels = db.labels_
# number of clusters in labels, ignoring noise if present
n_clusters_ = len(set(labels)) - (1 if -1 in labels else 0)
print 'Number of clusters: %d' % n_clusters_
print 'Homogeneity: %0.3f' % metrics.homogeneity_score(labels_true, labels)
print 'Completeness: %0.3f' % metrics.completeness_score(labels_true, labels)
print 'V-meassure: %0.3f' % metrics.v_measure_score(labels_true, labels)
print 'Adjusted Rand Index: %0.3f' % metrics.adjusted_rand_score(labels_true, labels)
print 'Adjusted Mutual Information: %0.3f' % metrics.adjusted_mutual_info_score(labels_true, labels)
print 'Silhouette Coefficient: %0.3f' % metrics.silhouette_score(D, labels, metric='precomputed')
def test_KMeans_scores(self):
digits = datasets.load_digits()
df = pdml.ModelFrame(digits)
scaled = pp.scale(digits.data)
df.data = df.data.pp.scale()
self.assert_numpy_array_almost_equal(df.data.values, scaled)
clf1 = cluster.KMeans(init='k-means++', n_clusters=10,
n_init=10, random_state=self.random_state)
clf2 = df.cluster.KMeans(init='k-means++', n_clusters=10,
n_init=10, random_state=self.random_state)
clf1.fit(scaled)
df.fit_predict(clf2)
expected = m.homogeneity_score(digits.target, clf1.labels_)
self.assertEqual(df.metrics.homogeneity_score(), expected)
expected = m.completeness_score(digits.target, clf1.labels_)
self.assertEqual(df.metrics.completeness_score(), expected)
expected = m.v_measure_score(digits.target, clf1.labels_)
self.assertEqual(df.metrics.v_measure_score(), expected)
expected = m.adjusted_rand_score(digits.target, clf1.labels_)
self.assertEqual(df.metrics.adjusted_rand_score(), expected)
expected = m.homogeneity_score(digits.target, clf1.labels_)
self.assertEqual(df.metrics.homogeneity_score(), expected)
expected = m.silhouette_score(scaled, clf1.labels_, metric='euclidean',
sample_size=300, random_state=self.random_state)
result = df.metrics.silhouette_score(metric='euclidean', sample_size=300,
random_state=self.random_state)
self.assertAlmostEqual(result, expected)
def optimal_cutoff(Y,dist_mat,min_size):
labels = np.array([sch.fcluster(Y,c,criterion='distance') for c in Y[:,2]])
score = np.array([metrics.silhouette_score(dist_mat,l) for l in labels[:-min_size]])
c = Y[:-min_size,2]
f = interp(c,-score,kind='linear')
opt_c = opt.fmin(f,x0=c[2*min_size])
return opt_c
def getBestGMM(X,n_components,cv_types):
'''
Function that finds the best GMM cluster trying different gaussians and
different number of clusters
'''
lowest_bic = np.infty
bic = []
silhouette = []
for cv_type in cv_types:
for n_components in n_components_range:
# Fit a mixture of Gaussians with EM
gmm = mixture.GMM(n_components=n_components, covariance_type=cv_type)
gmm.fit(X)
bic.append(gmm.bic(X))
Y_predicted=gmm.predict(X)
if cv_type =='tied':
silhouette.append(metrics.silhouette_score(X, Y_predicted, metric='euclidean'))
#I only save the values for tied, because i know from the first run that its the best gaussian
if n_components>=1:
if bic[-1] < lowest_bic:
lowest_bic = bic[-1]
best_gmm = gmm
bic = np.array(bic)
color_iter = itertools.cycle(['k', 'r', 'g', 'b', 'c', 'm','y'])
return best_gmm,color_iter,bic,silhouette
def kmean_clusters(min_list, d, show, thresh):
np.random.seed(0)
k_rng = range(1, len(min_list))
est = [KMeans(n_clusters=k).fit(d) for k in k_rng]
silhouette_score = [
metrics.silhouette_score(d, e.labels_, metric='euclidean') for e in est[1:]]
within_sum_squares = [e.inertia_ for e in est]
diff_sq = [sq / within_sum_squares[0] for sq in within_sum_squares]
diff_sq_pd = pd.Series(diff_sq)
k_list = list(k_rng)
select_k = k_list[len(k_list) - 1]
thresh_pd = diff_sq_pd[diff_sq_pd < thresh]
if thresh_pd.shape[0] > 0:
select_k = k_list[thresh_pd.index[0]]
if show:
TLineDrawer.plot_elow_k_choice(k_rng, silhouette_score, within_sum_squares, select_k)
select_est = est[select_k - 1]
y_kmean = select_est.predict(d)
return y_kmean, select_k
def findClusterSize(data): #also returns the value of the Silhouette Coefficient (I trust it more than elbow method) K = range(2,10) meandistortions = [] silCoeffs = [] for k in K: kmeans = KMeans(n_clusters=k) kmeans.fit(data) meandistortions.append(sum(np.min(cdist(data,kmeans.cluster_centers_,'euclidean'),axis=1))/data.shape[0]) silCoeffs.append(metrics.silhouette_score(data,labels=kmeans.labels_,metric='euclidean')) print "\n\nMean Distortions" print "------------------" for i in K: print 'K:',i,'\t',meandistortions[i-2],'\t Silhouette Coeff: ',silCoeffs[i-2] kDistTuple = [] print "\nDistortion Decline" print "-----------------------" for i in range(1,len(meandistortions)): difference = meandistortions[i-1]-meandistortions[i] kDistTuple.append([difference,i+1]) for i in range(len(kDistTuple)-1): print 'K: %2d -> %2d %10.5f'%(kDistTuple[i][1],kDistTuple[i+1][1],kDistTuple[i][0]) kDistTuple.sort() print "\n\nElbow Method Suggestion for Clusters: ",kDistTuple[0][1] kSilCoeff = zip(silCoeffs,K) kSilCoeff.sort() print "Best Silhouette Coeff and cluster size: ",kSilCoeff[-1] return kSilCoeff[-1]
def compute_silhouette_score(X, tree, metric_measure):
'''
n : sample sizes |X|
num of clusters, k = [1..n]
for each value of k
P_k: partition of X having k cluster (based on the maximum distance (or the radius) of a cluster)
compute silhouette score for P_k
input: X : data
tree: ward tree
matric_measure ('euclidean', ...)
output: float array 1D size n
value of silhouette score of partion P_k
'''
n = len(X)
score = np.zeros(n-1)
print 'Length : ', n
for i in range(n-1):
#canot calculate the silhouette score for only one cluster
#should start from 2 clusters
k = i + 2
print '\n Cutting at k = ', k
label = _hc_cut(k,tree.children_, tree.n_leaves_)
print '\n Compute score ...'
s = metrics.silhouette_score(X, label, metric = metric_measure)
#s = silhouette_score_block(X, label, metric= metric_measure , sample_size=None )
score[k-2] = s
def fit_dbscan(data, eps, min_samples, normalize=True,
show=True, juxta_cluster_indices_grouped=None, threshold_legend=None):
X = np.transpose(data)
if normalize:
from sklearn.preprocessing import minmax_scale
minmax_scale(X, feature_range=(-1, 1), axis=0, copy=False)
from sklearn.cluster import DBSCAN
from sklearn import metrics
db = DBSCAN(eps=eps, min_samples=min_samples).fit(X)
core_samples_mask = np.zeros_like(db.labels_, dtype=bool)
core_samples_mask[db.core_sample_indices_] = True
labels = db.labels_
# Number of clusters in labels, ignoring noise if present.
n_clusters_ = len(set(labels)) - (1 if -1 in labels else 0)
score = metrics.silhouette_score(X, labels, sample_size=5000)
print('For eps={}, min_samples={}, estimated number of clusters={}'.format(eps, min_samples, n_clusters_))
print("Silhouette Coefficient: {}".format(score))
if show:
pf.show_clustered_tsne(db, X, juxta_cluster_indices_grouped, threshold_legend)
return db, n_clusters_, labels, core_samples_mask, score
def calculateNumberOfIdealClusters(maxAmount, corpus):
print "Initializing silhouette analysis"
range_n_clusters = range(2, maxAmount) # max amount of clusters equal to amount of jobs
silhouette_high = 0;
silhouette_high_n_clusters = 2;
for n_clusters in range_n_clusters:
# Initialize the clusterer with n_clusters value
cluster = AgglomerativeClustering(n_clusters=n_clusters, linkage="ward", affinity="euclidean")
cluster_labels = cluster.fit_predict(corpus)
# The silhouette_score gives the average value for all the samples.
# This gives a perspective into the density and separation of the formed clusters
silhouette_avg = silhouette_score(corpus, cluster_labels)
print "For n_clusters = %d, the average silhouette_score is: %.5f" % (n_clusters, silhouette_avg)
if (silhouette_avg > silhouette_high):
silhouette_high = silhouette_avg
silhouette_high_n_clusters = n_clusters
# Compute the silhouette scores for each sample
sample_silhouette_values = silhouette_samples(corpus, cluster_labels)
print ("Highest score = %f for n_clusters = %d" % (silhouette_high, silhouette_high_n_clusters))
return silhouette_high_n_clusters
def _cluster(params):
cls = None
method = sh.getConst('method')
if method=='kmedoid':
assert False
# from kmedoid import kmedsoid
# cls = kmedoid
elif method=='dbscan':
from sklearn.cluster import DBSCAN
cls = DBSCAN(eps=params['eps'],min_samples=params['min_samples'],
metric='precomputed')
else:
assert False, 'FATAL: unknown cluster method'
##
mat = sh.getConst('mat')
labels = cls.fit_predict(mat)
nLabels = len(set(labels))
##
sil = None; cal = None
if (nLabels >= 2)and(nLabels <= len(labels)-1):
sil = met.silhouette_score(mat,labels,'precomputed')
cal = met.calinski_harabaz_score(mat,labels)
perf = dict(silhouette_score=sil,calinski_harabaz_score=cal)
return (labels,perf)
pca = PCA(n_components=0.95)
X_pca = pca.fit_transform(cc_clean)
# Kmeans for 2 to 8 clusters
KS = range(2, 10)
# storage
inertia = []
silo = []
for k in KS:
km = KMeans(k)
km.fit(X_pca)
labs = km.predict(X_pca)
inertia.append(km.inertia_)
silo.append(silhouette_score(X_pca, labs))
print(silo)
# plot
plt.figure(figsize=(15,5))
plt.subplot(1, 2, 1)
plt.title("Inertia")
sns.lineplot(KS, inertia)
plt.subplot(1, 2, 2)
plt.title("Silohouette Score")
sns.lineplot(KS, silo)
for i in range(49):
cluster_num = i + 2
print('the cluster num is: %d' % cluster_num)
# Set the K-Means input vector here #
Kmeans = KMeans(n_clusters=cluster_num,
random_state=None).fit(costKmeans_norm)
labels = np.array(Kmeans.labels_)
BC, WC = My_harabaz_score(costKmeans_norm, Kmeans.labels_, cluster_num)
temp_score = metrics.calinski_harabaz_score(costKmeans_norm,
Kmeans.labels_)
print('the harabaz score is: %f' % temp_score)
harabaz_score.append(temp_score)
temp_score1 = metrics.silhouette_score(costKmeans_norm,
Kmeans.labels_,
metric='euclidean')
sil_coe.append(temp_score1)
print('the sil_coe is: %f' % temp_score1)
BC_score.append(BC)
WC_score.append(WC)
print(BC)
print(WC)
print(labels)
labelsforKmeans.append(labels)
sil_coe_diff = np.diff(sil_coe)
harabaz_score_diff = np.diff(harabaz_score)
centers = pca.transform(clusterer.means_)
figname = create_path("fig",
sys.argv[1],
"GMM",
sys.argv[2],
filename=("%d_%s_gmm_vis.png" %
(n_clusters, covariance_type)))
visualize_cluster(X_vis, cluster_labels, n_clusters, centers,
figname)
ari = metrics.adjusted_rand_score(y, cluster_labels)
ami = metrics.adjusted_mutual_info_score(y, cluster_labels)
nmi = metrics.normalized_mutual_info_score(y, cluster_labels)
fms = metrics.fowlkes_mallows_score(y, cluster_labels)
sil = metrics.silhouette_score(X,
cluster_labels,
metric='euclidean')
chi = metrics.calinski_harabaz_score(X, cluster_labels)
dbi = metrics.davies_bouldin_score(X, cluster_labels)
print("Adjusted Rand index: %.6f" % ari)
print("Adjusted Mutual Information: %.6f" % ami)
print("Normalized Mutual Information: %.6f" % nmi)
print("Fowlkes-Mallows score: %.6f" % fms)
print("Silhouette Coefficient: %.6f" % sil)
print("Calinski-Harabaz Index: %.6f" % chi)
print("Davies-Bouldin Index: %.6f" % dbi)
ari_score.append(ari)
ami_score.append(ami)
nmi_score.append(nmi)
core_samples = db.core_sample_indices_
labels = db.labels_
# Number of clusters in labels, ignoring noise if present.
n_clusters_ = len(set(labels)) - (1 if -1 in labels else 0)
print('Estimated number of clusters: %d' % n_clusters_)
print("Homogeneity: %0.3f" % metrics.homogeneity_score(labels_true, labels))
print("Completeness: %0.3f" % metrics.completeness_score(labels_true, labels))
print("V-measure: %0.3f" % metrics.v_measure_score(labels_true, labels))
print("Adjusted Rand Index: %0.3f"
% metrics.adjusted_rand_score(labels_true, labels))
print("Adjusted Mutual Information: %0.3f"
% metrics.adjusted_mutual_info_score(labels_true, labels))
print("Silhouette Coefficient: %0.3f"
% metrics.silhouette_score(X, labels))
##############################################################################
# Plot result
import pylab as pl
# Black removed and is used for noise instead.
unique_labels = set(labels)
colors = pl.cm.Spectral(np.linspace(0, 1, len(unique_labels)))
for k, col in zip(unique_labels, colors):
if k == -1:
# Black used for noise.
col = 'k'
markersize = 6
class_members = [index[0] for index in np.argwhere(labels == k)]
cluster_core_samples = [index for index in core_samples
threedee.set_ylabel('y')
threedee.set_zlabel('z')
fig_2 = threedee.get_figure()
fig_2.suptitle('3D Clustered Data', fontsize=16)
fig_2.savefig('Q6_1.png')
plt.show()
#given that we know/can know the averages of these can also work an error from this
#and see if this is a better indicator
#compare the created samples mean and the predicted means
means_3 = np.array(clf_3.means_)
covariance_3 = clf_3.covariances_
print("The silhouette score is "+ str(silhouette_score(gen_df, labels_3, metric = 'euclidean')))
print("examining the means of the generated GMMs and the predicted ones")
print("counter mean")
print(counter_mean)
print("predicted mean")
print(means_3)
mean_difference = means_3 - counter_mean
print("The differences between the means in this model")
print(mean_difference)
print("This is because the large values dominate the mean")
#Large values are over represented in the mean
#The problem here is that there is far too much overlap in these distributions
#so while the model can fit it the means it provides mean nothing
#want to test that the model would work if the distribution were spread out
elapsed = timeit.default_timer() - start_time
print('Execution time: {0:.4f} sec'.format(elapsed))
x, y = zip(*sorted(
sil_coef.items())) # unpack a list of pairs into two tuples
plt.plot(x, y)
plt.xlabel('Number of Clusters')
plt.ylabel('Silhouette Score')
plt.show()
return sil_coef
sil_coef = cluster(df_N, 10)
#%%
print('Estimated number of clusters: %d' % n_clusters_)
print("Homogeneity: %0.3f" % metrics.homogeneity_score(phi_true, phi_predict))
print("Completeness: %0.3f" %
metrics.completeness_score(phi_true, phi_predict))
print("V-measure: %0.3f" % metrics.v_measure_score(phi_true, phi_predict))
print("Adjusted Rand Index: %0.3f" %
metrics.adjusted_rand_score(phi_true, phi_predict))
print("Adjusted Mutual Information: %0.3f" %
metrics.adjusted_mutual_info_score(phi_true, phi_predict))
print("Silhouette Coefficient: %0.3f" %
metrics.silhouette_score(dfHeurN, phi_predict, metric='sqeuclidean'))
def silhouette_coefficient(dataSet):
# List of number of clusters
range_n_clusters = [2, 3, 4, 5, 6]
X = dataSet
# pca = decomposition.PCA(n_components=2)
# pca.fit(X)
# X = pca.transform(X)
# For each number of clusters, perform Silhouette analysis and visualize the results.
for n_clusters in range_n_clusters:
# Perform k-means.
kmeans = KMeans(n_clusters=n_clusters, random_state=10)
y_pred = kmeans.fit_predict(X)
# Compute the cluster homogeneity and completeness.
homogeneity = metrics.homogeneity_score(y_pred, y_pred)
completeness = metrics.completeness_score(y_pred, y_pred)
# Compute the Silhouette Coefficient for each sample.
s = metrics.silhouette_samples(X, y_pred)
# Compute the mean Silhouette Coefficient of all data points.
s_mean = metrics.silhouette_score(X, y_pred)
# For plot configuration -----------------------------------------------------------------------------------
fig, (ax1, ax2) = plt.subplots(1, 2)
fig.set_size_inches(18, 7)
# Configure plot.
plt.suptitle(
'Silhouette analysis for K-Means clustering with n_clusters: {}'.
format(n_clusters),
fontsize=14,
fontweight='bold')
# Configure 1st subplot.
ax1.set_title('Silhouette Coefficient for each sample')
ax1.set_xlabel("The silhouette coefficient values")
ax1.set_ylabel("Cluster label")
ax1.set_xlim([-1, 1])
ax1.set_ylim([0, len(X) + (n_clusters + 1) * 10])
# Configure 2st subplot.
ax2.set_title(
'Homogeneity: {}, Completeness: {}, Mean Silhouette score: {}'.
format(homogeneity, completeness, s_mean))
ax2.set_xlabel("Feature space for the 1st feature")
ax2.set_ylabel("Feature space for the 2nd feature")
# For 1st subplot ------------------------------------------------------------------------------------------
# Plot Silhouette Coefficient for each sample
y_lower = 10
for i in range(n_clusters):
ith_s = s[y_pred == i]
ith_s.sort()
size_cluster_i = ith_s.shape[0]
y_upper = y_lower + size_cluster_i
color = cm.spectral(float(i) / n_clusters)
ax1.fill_betweenx(np.arange(y_lower, y_upper),
0,
ith_s,
facecolor=color,
edgecolor=color,
alpha=0.7)
ax1.text(-0.05, y_lower + 0.5 * size_cluster_i, str(i))
y_lower = y_upper + 10
# Plot the mean Silhouette Coefficient using red vertical dash line.
ax1.axvline(x=s_mean, color="red", linestyle="--")
# For 2st subplot -------------------------------------------------------------------------------------------
#pca = decomposition.PCA(n_components=2)
#pca.fit(X)
#plot_X = pca.transform(X)
# Plot the predictions
colors = cm.spectral(y_pred.astype(float) / n_clusters)
ax2.scatter(X[:, 0], X[:, 1], c=colors)
return X
centers = kmeans.cluster_centers_
score = kmeans.score(df_scaled)
# Compute Clustering Metrics
n_clusters_ = len(centers)
print('Number of clusters: %d' % n_clusters_)
#print("Homogeneity: %0.3f" % metrics.homogeneity_score(phi_true, phi_predict))
#print("Completeness: %0.3f" % metrics.completeness_score(phi_true, phi_predict))
#print("V-measure: %0.3f" % metrics.v_measure_score(phi_true, phi_predict))
#print("Adjusted Rand Index: %0.3f"
# % metrics.adjusted_rand_score(phi_true, phi_predict))
#print("Adjusted Mutual Information: %0.3f"
# % metrics.adjusted_mutual_info_score(phi_true, phi_predict))
print("Silhouette Coefficient: %0.3f" %
metrics.silhouette_score(df_scaled, phi_predict, metric='sqeuclidean'))
# timeit statement
elapsed = timeit.default_timer() - start_time
#%%
import math
df = df_metrics
plt.scatter(df.AvgStart, df.AvgEnergy)
#%%
def round10(x):
return int(math.ceil(x / 5.0)) * 5
def main():
dataset = pd.read_csv('dataset.csv')
positive = dataset.loc[dataset['Lab Status'] == 'Positive ID']
latitude = get_column_value(positive, 'Latitude').tolist()
longitude = get_column_value(positive, 'Longitude').tolist()
date = get_column_value(positive, 'Detection Date').tolist()
date = pd.to_datetime(date)
interval = (date - date[0]).days
interval = interval - np.min(interval)
data = []
for i, la in enumerate(latitude):
data.append([latitude[i], longitude[i], interval[i]])
data = np.array(data)
data = data[np.argsort(data[:, 2])]
data_scale = preprocessing.scale(data)
SSE = []
for k in range(2, 9):
kmeans = KMeans(n_clusters=k, random_state=0).fit(data_scale)
SSE.append(kmeans.inertia_)
X = range(2, 9)
plt.xlabel('Number of Clusters(k)')
plt.ylabel('SSE')
plt.title('SSE vs k')
plt.plot(X, SSE, 'o-')
plt.show()
Scores = []
for k in range(2, 9):
kmeans = KMeans(n_clusters=k, random_state=0).fit(data)
Scores.append(
silhouette_score(data, kmeans.labels_, metric='euclidean'))
X = range(2, 9)
plt.xlabel('Number of Clusters(k)')
plt.ylabel('Silhouette Coefficient')
plt.title('Silhouette Coefficient vs k')
plt.plot(X, Scores, 'o-')
plt.show()
cluster_num = 3
kmeans = KMeans(n_clusters=cluster_num, random_state=0).fit(data_scale)
label = kmeans.labels_
centers = []
label_list = []
for i in range(cluster_num):
label_list.append(data[label == i, 0:2].tolist())
centers.append(np.mean(data[label == i], axis=0).tolist())
centers = np.array(centers)
centers_list = np.delete(centers, -1, axis=1).tolist()
centers = centers[np.argsort(centers[:, 2])]
print(centers)
ax1 = plt.axes(projection='3d')
ax1.scatter3D(data[:, 1],
data[:, 0],
data[:, 2],
c=kmeans.labels_,
cmap='rainbow')
ax1.scatter3D(centers[:, 1],
centers[:, 0],
centers[:, 2],
c='black',
s=150,
alpha=0.5)
plt.show()
x = centers[:, 1].reshape((-1, 1))
y = centers[:, 0]
reg = LinearRegression().fit(x, y)
k = reg.coef_[0]
b = reg.intercept_
print("Y = %.5fX + (%.5f)" % (k, b))
plt.scatter(data[:, 1], data[:, 0], c=label, cmap='rainbow')
plt.scatter(centers[:, 1], centers[:, 0], c='black', s=150, alpha=0.5)
data = data[np.argsort(data[:, 1])]
plt.plot(data[np.argsort(data[:, 1])][:, 1].reshape((-1, 1)),
reg.predict(data[np.argsort(data[:, 1])][:, 1].reshape((-1, 1))),
c='b',
linestyle='--')
plt.xlabel('Longitude')
plt.ylabel('Latitude')
plt.title('Linear Regression of Cluster Centers(k=%d)' % cluster_num)
plt.grid()
plt.show()
cluster_foot_x, cluster_foot_y = get_foot_point(centers[-1, 1],
centers[-1, 0], k, b)
print("center-%d distance to line:%.5f" %
(cluster_num,
get_distance([centers[-1, 1], centers[-1, 0]],
[cluster_foot_x, cluster_foot_y])))
sum_dis = 0
for i in range(data.shape[0]):
foot_x, foot_y = get_foot_point(data[i, 1], data[i, 0], k, b)
sum_dis += get_distance([data[i, 1], data[i, 0]], [foot_x, foot_y])
print("sum_dis:%.5f" % sum_dis)
colors = ['blue', 'green', 'orange', 'pink', 'purple', 'red']
map = folium.Map(location=[48.9938, -122.702],
zoom_start=8,
tiles="OpenStreetMap")
for i in range(len(label_list)):
point_list = label_list[i]
for point in range(len(point_list)):
folium.CircleMarker(radius=2.5,
location=label_list[i][point],
color=colors[i],
fill=True,
fill_color=colors[i],
fill_opacity=1).add_to(map)
for i in range(len(centers_list)):
folium.CircleMarker(cradius=6,
location=centers_list[i],
color=colors[i],
fill=True,
fill_color=colors[i],
fill_opacity=0.3).add_to(map)
map.save('map_cluster%d.html' % cluster_num)
labels = d.labels_
# Number of clusters in labels, ignoring noise if present.
n_clusters_ = len(set(labels)) - (1 if -1 in labels else 0)
n_noise_ = list(labels).count(-1)
print('Estimated number of clusters: %d' % n_clusters_)
print('Estimated number of noise points: %d' % n_noise_)
print("Homogeneity: %0.3f" % metrics.homogeneity_score(labels_true, labels))
print("Completeness: %0.3f" % metrics.completeness_score(labels_true, labels))
print("V-measure: %0.3f" % metrics.v_measure_score(labels_true, labels))
print("Adjusted Rand Index: %0.3f" %
metrics.adjusted_rand_score(labels_true, labels))
print("Adjusted Mutual Information: %0.3f" %
metrics.adjusted_mutual_info_score(labels_true, labels))
print("Silhouette Coefficient: %0.3f" % metrics.silhouette_score(X, labels))
print(db.core_sample_indices_)
print(db.labels_ == d.labels_)
if (db.labels_ == d.labels_).all():
print('lists the same')
else:
print('lists differ')
print('self.x from Dbscan')
print(d.x)
# #############################################################################
# Plot result
import matplotlib.pyplot as plt
# Black removed and is used for noise instead.
unique_labels = set(labels)
colors = [
def silhouette_elbow__analysis(X, range_n_clusters, all_cluster_labels, all_centers):
"""
:param X: 原始样本
:param range_n_clusters: K的取值情况， list
:param all_cluster_labels: 簇标签结果 list
:param all_centers: 每种K值情况下的簇中心 list
:return:
"""
assert len(all_cluster_labels) == len(all_centers) == len(range_n_clusters)
plt.figure(figsize=(10, 8))
row_plot = 3 # 子图的行数
all_dist = []
for n, n_clusters in enumerate(range_n_clusters):
# ================= 轮廓分析法 ============================
cluster_labels = all_cluster_labels[n]
plt.subplot(row_plot, (len(range_n_clusters) + 1) // row_plot, n + 1)
plt.xlim([-0.1, 1]) # 设置x轴的范围（轮廓系数）
plt.ylim([0, len(X) + (n_clusters + 1) * 10]) # 顶端的间隙
silhouette_avg = silhouette_score(X, cluster_labels) # 所有样本的轮廓系数均值
print(" 当 n_clusters = ", n_clusters, "时，轮廓系数为: ", silhouette_avg)
# 计算每个样本对应的轮廓系数
sample_silhouette_values = silhouette_samples(X, cluster_labels)
y_lower = 10
for i in range(n_clusters): # 遍历每一个簇
# 取第i个簇中对应所有样本的轮廓系数，并进行排序
s_values = sample_silhouette_values[cluster_labels == i]
s_values.sort()
size_cluster_i = s_values.shape[0] # 得到第i个簇的样本数量
y_upper = y_lower + size_cluster_i # 图中每个簇在y轴上的宽度
# 限定y的范围，填充x1和x2所围成的区域
plt.fill_betweenx(y=np.arange(y_lower, y_upper), x1=0, x2=s_values, alpha=0.7)
# 在y轴右侧标记每个簇的序号
plt.text(-0.05, y_lower + 0.5 * size_cluster_i, str(i))
# 计算下一个条形图y轴的其实值
y_lower = y_upper + 10 # 10 for the 0 samples
fm.fontManager.addfont('../data/SimHei.ttf')
plt.rcParams['font.sans-serif'] = ['SimHei'] # 用来 正常显示中文标签
plt.title(f"K = {n_clusters} 时的轮廓系数图", fontsize=12)
plt.xlabel("轮廓系数", fontsize=12)
plt.ylabel("聚类簇序号", fontsize=12)
# 以x=silhouette_avg 画一条平行于y轴的线
plt.axvline(x=silhouette_avg, color="red", linestyle="--")
plt.yticks([]) # 去掉y轴的刻度
plt.xticks([-0.1, 0, 0.2, 0.4, 0.6, 0.8, 1]) # 设置x轴的刻度
# ============ 肘部法计算簇内距离和并可视化 =========================
dist = 0
centers = all_centers[n]
for i in range(n_clusters): # 遍历每一个簇，计算当前簇的簇内距离
x_data = X[cluster_labels == i]
tmp = np.sum((x_data - centers[i]) ** 2, axis=1)
dist += np.sum(np.sqrt(tmp)) # 累计当前聚类结果下所有簇的簇内距离和
all_dist.append(dist)
plt.subplot(row_plot, (len(range_n_clusters) + 1) // row_plot, len(range_n_clusters) + 1)
plt.title("肘部法结果")
plt.plot(range_n_clusters, all_dist) # 绘制肘部曲线
plt.scatter(range_n_clusters, all_dist) # 绘制各个点
for i in range(len(range_n_clusters)): # 在图上进行K值标记
plt.annotate(f"k = {range_n_clusters[i]}",
xy=(range_n_clusters[i], all_dist[i]), fontsize=14,
xytext=(range_n_clusters[i] + 0.1, all_dist[i]))
plt.hlines(all_dist[i], xmin=0, xmax=range_n_clusters[i], color="red", linestyle="--")
plt.xlim(range_n_clusters[0] - 0.5, range_n_clusters[-1] + 0.8) # 调整范围
plt.ylim(all_dist[-1] * 0.9, all_dist[0] + all_dist[-1] * 0.1)
plt.yticks([]) # 去掉y轴上的刻度显示
plt.xlabel("K", fontsize=12)
plt.ylabel("distance", fontsize=12)
plt.tight_layout()
plt.show()
def getSilhouetteCoeff(data, labels):
silCoeff = silhouette_score(data, labels, metric='euclidean')
return silCoeff
def my_score(X, y):
return mutual_info_classif(X, y, random_state=724)
selectedfeatures = SelectKBest(my_score, k=25)
selectedfeatures.fit(data_train, label_train)
smalldata_train = selectedfeatures.transform(
data_train) #ti qu guo te zheng de 25 wei xun lian shu ju
#kmeans = KMeans(n_clusters=50, random_state=0).fit_predict(data_train)
smalldata_train = np.hstack(
(smalldata_train, data_train[0:, original_length:]))
best_silhouette_score = -2
for nclusters in range(40, 120, 20):
kmeans_model = KMeans(n_clusters=nclusters, random_state=926)
kmeans_model.fit(smalldata_train)
tmp = silhouette_score(smalldata_train,
kmeans_model.labels_,
sample_size=40000)
print("n score: %0.3f %0.3f" % (nclusters, tmp))
if (tmp > best_silhouette_score):
best_silhouette_score = tmp
best_n_clusters = nclusters
print("%0.3f best_n_clusters" % (best_n_clusters))
kmeans_model = KMeans(n_clusters=best_n_clusters, random_state=926)
kmeans_model.fit(smalldata_train)
kmeans = kmeans_model.labels_
# print kmeans
#print>>f, kmeans.size
split = np.zeros(10000)
nosplit = np.zeros(10000)
clusters = 0
len1 = int(data_train.size /
def vis_cluster(X,X_pca,n_clusters):
y_lower = 10 # y값의 기준
kmeans = KMeans(n_clusters= n_clusters, random_state=42)
cluster_labels = kmeans.fit_predict(X_pca)
# clustering
silhouette_avg = silhouette_score(X, cluster_labels)
sample_silhouette_values = silhouette_samples(X, cluster_labels)
# visual
fig, (ax1, ax2) = plt.subplots(1, 2)
fig.set_size_inches(18, 7)
# The 1st subplot is the silhouette plot
# The silhouette coefficient can range from -1, 1 but in this example all
# lie within [-0.1, 1]
ax1.set_xlim([-0.1, 1])
# The (n_clusters+1)*10 is for inserting blank space between silhouette
# plots of individual clusters, to demarcate them clearly.
ax1.set_ylim([0, len(X) + (n_clusters + 1) * 10])
for i in range(n_clusters):
# Aggregate the silhouette scores for samples belonging to
# cluster i, and sort them
ith_cluster_silhouette_values = \
sample_silhouette_values[cluster_labels == i]
ith_cluster_silhouette_values.sort()
size_cluster_i = ith_cluster_silhouette_values.shape[0]
y_upper = y_lower + size_cluster_i
color = cm.nipy_spectral(float(i) / n_clusters)
ax1.fill_betweenx(np.arange(y_lower, y_upper),
0, ith_cluster_silhouette_values,
facecolor=color, edgecolor=color, alpha=0.7)
# Label the silhouette plots with their cluster numbers at the middle
ax1.text(-0.05, y_lower + 0.5 * size_cluster_i, str(i))
# Compute the new y_lower for next plot
y_lower = y_upper + 10 # 10 for the 0 samples
ax1.set_title("The silhouette plot for the various clusters.")
ax1.set_xlabel("The silhouette coefficient values")
ax1.set_ylabel("Cluster label")
# The vertical line for average silhouette score of all the values
ax1.axvline(x=silhouette_avg, color="red", linestyle="--")
ax1.set_yticks([]) # Clear the yaxis labels / ticks
ax1.set_xticks([-0.1, 0, 0.2, 0.4, 0.6, 0.8, 1])
# 2nd Plot showing the actual clusters formed
colors = cm.nipy_spectral(cluster_labels.astype(float) / n_clusters)
ax2.scatter(X_pca[:, 0], X_pca[:, 1], marker='.', s=30, lw=0, alpha=0.7,
c=colors, edgecolor='k')
# Labeling the clusters
centers = kmeans.cluster_centers_
# Draw white circles at cluster centers
ax2.scatter(centers[:, 0], centers[:, 1], marker='o',
c="white", alpha=1, s=200, edgecolor='k')
for i, c in enumerate(centers):
ax2.scatter(c[0], c[1], marker='$%d$' % i, alpha=1,
s=50, edgecolor='k')
ax2.set_title("The visualization of the clustered data.")
ax2.set_xlabel("Feature space for the 1st feature")
ax2.set_ylabel("Feature space for the 2nd feature")
plt.suptitle(("Silhouette analysis for KMeans clustering on sample data "
"with n_clusters = %d" % n_clusters),
fontsize=14, fontweight='bold')
plt.show()
clusters = fcluster(distance, k, criterion='maxclust')
plt.figure(figsize=(10, 8))
plt.scatter(x[:, 0], x[:, 1], c=clusters, cmap='prism')
'''K-Means'''
from sklearn.cluster import KMeans
model = KMeans(n_clusters=5)
model.fit(x)
y_kmeans = model.predict(x)
plt.scatter(x[:, 0], x[:, 1], c=y_kmeans, s=10, cmap='inferno')
centers = model.cluster_centers_
plt.scatter(centers[:, 0], centers[:, 1], c='cyan', s=300)
from mlxtend.plotting import plot_decision_regions
print(model.inertia_)
elbow = []
for i in range(1, 15):
kmeans = KMeans(n_clusters=i).fit(x)
elbow.append([i, kmeans.inertia_])
plt.plot(pd.DataFrame(elbow)[0], pd.DataFrame(elbow)[1])
from sklearn.metrics import silhouette_score
silhoutte = []
for i in range(2, 8):
kmeans = KMeans(n_clusters=i).fit(x)
silhoutte.append([i, silhouette_score(x, kmeans.labels_)])
plt.plot(pd.DataFrame(silhoutte)[0], pd.DataFrame(silhoutte)[1])
model.fit(X)
# append model to cluster list
clusters.append(model)
inertia_vals.append(model.inertia_)
# plot the inertia vs K values
plt.plot(range(2, 15, 1), inertia_vals, marker='*')
plt.show()
from sklearn.metrics import silhouette_score
#
# print(clusters[1])
# print("Silhouette score for k=4", silhouette_score(X, clusters[1].predict(X)))
print(clusters[2])
print("Silhouette score for k=4", silhouette_score(X, clusters[2].predict(X)))
print(clusters[3])
print("Silhouette score for k=5", silhouette_score(X, clusters[3].predict(X)))
print(clusters[4])
print("Silhouette score for k=6", silhouette_score(X, clusters[4].predict(X)))
print(clusters[5])
print("Silhouette score for k=7", silhouette_score(X, clusters[5].predict(X)))
# K means clustering using the term vector
kmeans = KMeans(n_clusters=6, random_state=rs).fit(X)
# function to visualise text cluster. Useful for the assignment too
# In[9]:
from sklearn.cluster import KMeans
Cluster = KMeans(n_clusters=3, random_state=2)
Cluster.fit(data)
y_pred = Cluster.predict(data)
plt.scatter(data_arr[:, 0], data_arr[:, 1], c=y_pred, s=50, cmap='plasma')
plt.rcParams.update({'figure.figsize': (10, 7.5), 'figure.dpi': 100})
# In[10]:
Cluster.fit(data)
y_pred = Cluster.predict(test)
plt.scatter(test[:, 0], test[:, 1], c=y_pred, s=50, cmap='plasma')
plt.rcParams.update({'figure.figsize': (10, 7.5), 'figure.dpi': 100})
# In[11]:
from sklearn.metrics import silhouette_score
for i in range(2, 10):
clusterer = KMeans(n_clusters=i, random_state=i)
cluster_labels = clusterer.fit_predict(data)
silhouette_avg = silhouette_score(data, cluster_labels)
print("For n_clusters =", i, "The average silhouette_score is :",
silhouette_avg)
# In[ ]:
# The silhouette coefficient can range from -1, 1 but in this example all
# lie within [-0.1, 1]
ax1.set_xlim([-0.1, 1])
# The (n_clusters+1)*10 is for inserting blank space between silhouette
# plots of individual clusters, to demarcate them clearly.
ax1.set_ylim([0, len(X) + (n_clusters + 1) * 10])
# Initialize the clusterer with n_clusters value and a random generator
# seed of 10 for reproducibility.
clusterer = KMeans(n_clusters=n_clusters, random_state=10)
cluster_labels = clusterer.fit_predict(X)
# The silhouette_score gives the average value for all the samples.
# This gives a perspective into the density and separation of the formed
# clusters
silhouette_avg = silhouette_score(X, cluster_labels)
print("For n_clusters =", n_clusters,
"The average silhouette_score is :", silhouette_avg)
# Compute the silhouette scores for each sample
sample_silhouette_values = silhouette_samples(X, cluster_labels)
y_lower = 10
for i in range(n_clusters):
# Aggregate the silhouette scores for samples belonging to
# cluster i, and sort them
ith_cluster_silhouette_values = \
sample_silhouette_values[cluster_labels == i]
ith_cluster_silhouette_values.sort()
def tune_hyperparameters(data_list, if_tune_m=True, m_range=None, if_tune_dim=True, dim_range=None, if_tune_p=False, p_range=None, log_norm=True, l2_norm=True, true_labels=None, verbose=True):
# Specify data normalization
data_list = preprocess(data_list, log_norm=log_norm, l2_norm=l2_norm)
num_datasets = len(data_list)
# Impute m if None
if m_range==None:
m_est = max(m_estimate(data_list))
if if_tune_m:
m_range = [m_est+i*5 for i in range(-3, 3)]
else:
m_range = [m_est]
print('WARNING no value of m is given, default m={} for the dataset(s) from estimation.'.format(m_est))
# Impute dim if None
if dim_range==None:
dim_est = dim_estimate(data_list)
if if_tune_dim:
dim_range = [dim_est+i*10 for i in range(-2, 2)]
else:
dim_range = [dim_est]
print('WARNING no value of dim is given, default dim={} for the dataset(s) from estimation.'.format(dim_est))
# Impute p if None
if p_range==None:
if if_tune_p:
p_range = [0.1, 0.3, 0.5]
else:
p_range = [0.3]
print('WARNING no value of p is given, default p=0.3 for the dataset(s) from estimation.')
# If ground truth given, find n_clusters
if true_labels is not None:
n_clusters = len(np.unique(true_labels))
out = []
if verbose:
print('Testing hyperparameters in the range below:')
print('Range for m: {}'.format(m_range))
print('Range for dim: {}'.format(dim_range))
print('Range for p: {}'.format(p_range))
for m in m_range:
for n_dim in dim_range:
for p in p_range:
if m*p < 3:
print('Skip m={} and p={} as the number of ghost cells is smaller than 3.'.format(m, p))
continue
ZW = run_OCAT(data_list=data_list, m_list=[m]*num_datasets, dim=n_dim, p=p, log_norm=False, l2_norm=False)
if true_labels is None:
labels_pred, n_clusters = evaluate_clusters(ZW, return_num_cluster=True)
sil_score = silhouette_score(ZW, labels_pred)
out.append([m, n_dim, p, n_clusters, sil_score])
else:
labels_pred = evaluate_clusters(ZW, num_cluster=n_clusters)
NMI_cell = normalized_mutual_info_score(true_labels, labels_pred)
AMI_cell = adjusted_mutual_info_score(true_labels, labels_pred)
ARI_cell = adjusted_rand_score(true_labels, labels_pred)
out.append([m, n_dim, p, NMI_cell, AMI_cell, ARI_cell])
out = np.array(out)
if true_labels is not None:
df = pd.DataFrame(data=out, columns=['m', 'n_dim', 'p', 'NMI_score', 'AMI_score', 'ARI_score'])
else:
df = pd.DataFrame(data=out, columns=['m', 'n_dim', 'p', 'n_clusters', 'silhoutte_score'])
if verbose:
print(df)
return df
import matplotlib.pyplot as plt
for i in clases:
dlabels = np.where(dy[:, 0] == i)[0]
plt.plot(data1[dlabels, 0], data1[dlabels, 1], 'x')
plt.xlabel('Agrupamiento original')
plt.show()
# ### Silhoulette Score de los datos con la clasificacion original
#
# In[527]:
from sklearn.metrics import silhouette_score
silhouette_score(data1, dy[:, 0] - 1, metric='sqeuclidean')
# #### Clasifico Con las 2 mejores caracteristicas y grafico
# In[528]:
cm = cmeans(nclusters=clases.shape[0])
cm.fit(datas[0], m=3)
# ## Matriz de pertenencias de los datos a los clusters
# In[531]:
membership = cm.predict(datas[0], m=3)
np.round(membership, 1)
def illustration(data, range_n_clusters):
"""
TBD
"""
# Scale des données obligatoire avant la réduction des dimensions
std_scale = preprocessing.StandardScaler().fit(data)
X = std_scale.transform(data)
for n_clusters in range_n_clusters:
# Create a subplot with 1 row and 2 columns
fig, (ax1, ax2) = plt.subplots(1, 2)
fig.set_size_inches(36, 14)
# The 1st subplot is the silhouette plot
# The silhouette coefficient can range from -1, 1 but in this example all
# lie within [-0.1, 1]
ax1.set_xlim([-0.1, 1])
# The (n_clusters+1)*10 is for inserting blank space between silhouette
# plots of individual clusters, to demarcate them clearly.
ax1.set_ylim([0, len(X) + (n_clusters + 1) * 10])
# Initialize the clusterer with n_clusters value and a random generator
# seed of 10 for reproducibility.
clusterer = KMeans(n_clusters=n_clusters, random_state=10)
cluster_labels = clusterer.fit_predict(X)
# The silhouette_score gives the average value for all the samples.
# This gives a perspective into the density and separation of the formed
# clusters
silhouette_avg = silhouette_score(X, cluster_labels)
print("For n_clusters =", n_clusters,
"The average silhouette_score is :", silhouette_avg)
# Compute the silhouette scores for each sample
sample_silhouette_values = silhouette_samples(X, cluster_labels)
y_lower = 10
for i in range(n_clusters):
# Aggregate the silhouette scores for samples belonging to
# cluster i, and sort them
ith_cluster_silhouette_values = sample_silhouette_values[
cluster_labels == i]
ith_cluster_silhouette_values.sort()
size_cluster_i = ith_cluster_silhouette_values.shape[0]
y_upper = y_lower + size_cluster_i
color = cm.spectral(float(i) / n_clusters)
ax1.fill_betweenx(np.arange(y_lower, y_upper),
0,
ith_cluster_silhouette_values,
facecolor=color,
edgecolor=color,
alpha=0.7)
# Label the silhouette plots with their cluster numbers at the middle
ax1.text(-0.05, y_lower + 0.5 * size_cluster_i, str(i))
# Compute the new y_lower for next plot
y_lower = y_upper + 10 # 10 for the 0 samples
ax1.set_title("The silhouette plot for the various clusters.")
ax1.set_xlabel("The silhouette coefficient values")
ax1.set_ylabel("Cluster label")
# The vertical line for average silhouette score of all the values
ax1.axvline(x=silhouette_avg, color="red", linestyle="--")
ax1.set_yticks([]) # Clear the yaxis labels / ticks
ax1.set_xticks([-0.1, 0, 0.2, 0.4, 0.6, 0.8, 1])
# 2nd Plot showing the actual clusters formed
colors = cm.spectral(cluster_labels.astype(float) / n_clusters)
ax2.scatter(X[:, 0],
X[:, 1],
marker='.',
s=30,
lw=0,
alpha=0.7,
c=colors,
edgecolor='k')
# Labeling the clusters
centers = clusterer.cluster_centers_
# Draw white circles at cluster centers
ax2.scatter(centers[:, 0],
centers[:, 1],
marker='o',
c="white",
alpha=1,
s=200,
edgecolor='k')
for i, c in enumerate(centers):
ax2.scatter(c[0],
c[1],
marker='$%d$' % i,
alpha=1,
s=50,
edgecolor='k')
ax2.set_title("The visualization of the clustered data.")
ax2.set_xlabel("Feature space for the 1st feature")
ax2.set_ylabel("Feature space for the 2nd feature")
plt.suptitle(
("Silhouette analysis for KMeans clustering on sample data "
"with n_clusters = %d" % n_clusters),
fontsize=14,
fontweight='bold')
plt.show()
silhouette = {}
D = pairwise_distances(training_dataset, metric='euclidean')
for n_clusters in range_n_clusters:
print('number of clusters :{}'.format(n_clusters))
M, C = kmedoids.kMedoids(D, n_clusters)
labels = np.zeros(1173)
for label in C:
#print(label)
for point_idx in C[label]:
labels[point_idx] = label
silhouette_avg = silhouette_score(training_dataset, labels)
vcr_avg = calinski_harabasz_score(training_dataset, labels)
silhouette[n_clusters] = silhouette_avg
print("For n_clusters =", n_clusters, "The average silhouette_score is :",
silhouette_avg)
print("For n_clusters =", n_clusters, "The average silhouette_score is :",
vcr_avg)
#sample_silhouette_values = silhouette_samples(training_dataset, cluster_labels)
plt.figure()
plt.plot(list(inertia.keys()), list(inertia.values()))
plt.xlabel("Number of cluster")
plt.ylabel("SSE")
plt.ylim([0, 10])
plt.title('Instances')
plt.scatter(x1, x2)
colors = ['b', 'g', 'r', 'c', 'm', 'y', 'k', 'b']
markers = ['o', 's', 'D', 'v', '^', 'p', '*', '+']
clusters = [2, 3, 4, 5, 8]
subplot_counter = 1
sc_scores = []
for t in clusters:
subplot_counter += 1
plt.subplot(3, 2, subplot_counter)
kmeans_model = KMeans(n_clusters=t).fit(X)
for i, l in enumerate(kmeans_model.labels_):
plt.plot(x1[i], x2[i], color=colors[l], marker=markers[l], ls='None')
plt.xlim([0, 10])
plt.ylim([0, 10])
sc_score = silhouette_score(X, kmeans_model.labels_, metric='euclidean')
sc_scores.append(sc_score)
# 绘制轮廓系数与不同类簇数量的直观显示图。
plt.title('K = %s, silhouette coefficient= %0.03f' % (t, sc_score))
# 绘制轮廓系数与不同类簇数量的关系曲线。
plt.figure()
plt.plot(clusters, sc_scores, '*-')
plt.xlabel('Number of Clusters')
plt.ylabel('Silhouette Coefficient Score')
plt.show()
cah = AgglomerativeClustering(n_clusters=k)
cah.fit(centroids)
labels = cah.labels_
nv_centroids = pd.DataFrame(centroids)
nv_centroids["labels"] = labels
nv_centroids = nv_centroids.groupby("labels").mean()
# Consolidation du KMeans
clf_2 = KMeans(n_clusters=k, init=nv_centroids)
clf_2.fit(sv_data_scaled)
labels_final = clf_2.labels_
s_score = silhouette_score(sv_data_scaled,
labels_final,
metric="sqeuclidean")
s_scores.append(s_score)
plt.plot(k_clust, s_scores)
# Score de silhouette le plus élevé pour n_clusters = 2
# => meilleur nombre pour l'homogénéité intra-cluster et la séparation inter-cluster
# MAIS ne permet pas une bonne séparation inter-cluster, comme vu au dessus (van et bus très proche)
# Choisir 3 clusters semble donc pertinent quitte à avoir moins d'homogénéité intra-cluster
clf = KMeans(n_clusters=3)
clf.fit(sv_data_scaled)
labels = clf.labels_
init='k-means++',
max_iter=500,
n_init=1,
verbose=opts.verbose)
print("Clustering sparse data with %s" % km)
t0 = time()
km.fit(X)
print("done in %0.3fs" % (time() - t0))
print()
print("Homogeneity: %0.3f" % metrics.homogeneity_score(labels, km.labels_))
print("Completeness: %0.3f" % metrics.completeness_score(labels, km.labels_))
print("V-measure: %0.3f" % metrics.v_measure_score(labels, km.labels_))
print("Adjusted Rand-Index: %.3f" %
metrics.adjusted_rand_score(labels, km.labels_))
print("Silhouette Coefficient: %0.3f" %
metrics.silhouette_score(X, labels, sample_size=1000))
print()
if not (opts.n_components or opts.use_hashing):
print("Top terms per cluster:")
order_centroids = km.cluster_centers_.argsort()[:, ::-1]
terms = vectorizer.get_feature_names()
for i in range(true_k):
print("Cluster %d:" % i, end='')
for ind in order_centroids[i, :10]:
print(' %s' % terms[ind], end='')
print()
def get(self, x, labels):
return silhouette_score(x, labels, **self.kwargs)
print(i)
W = W[0:4]
biases= biases[0:4]
def encode(encoder_weights,encoder_biases,data):
res=data
for index, (w,b) in enumerate(zip(encoder_weights,encoder_biases)):
if index+1 == len(encoder_weights):
res= np.dot(res,w) + b
else:
res = np.maximum(0,np.dot(res,w) + b)
return res
res = encode(W,biases,X_test)
print(res.shape)
unique_labels = np.unique(y_test)
for index,unique_label in enumerate(unique_labels):
data_latent_space = res[y_test==unique_label]
plt.scatter(data_latent_space[:,0],data_latent_space[:,1],alpha=0.3,c =cmap(index))
plt.xlabel("Latent X")
plt.ylabel("Latest Y")
plt.title("Autoencoder results")
print(silhouette_score(res,y_test))
print("PCA silouette score( how good the clustering is made")
print(silhouette_score(res_pca,y_test))
def silhouette(pairWisePointDistance, clusterLabels):
# pairWisePointDistance: Array of pairwise distances between samples, or a feature array.
# clusterLabels: Predicted labels for each sample.
return silhouette_score(pairWisePointDistance, clusterLabels)
df_cluster[features] = scaler.fit_transform(df_cluster[features])
df_cluster.describe().transpose()
# Elbow Method: determine the appropriate number of K
inertias = {}
silhouettes = {}
for k in range(2, 11):
kmeans = KMeans(init='k-means++',
n_init=10,
n_clusters=k,
max_iter=1000,
random_state=42).fit(df_cluster)
inertias[
k] = kmeans.inertia_ # Inertia: Sum of distances of samples to their closest cluster center
silhouettes[k] = silhouette_score(df_cluster,
kmeans.labels_,
metric='euclidean')
plt.figure()
plt.grid(True)
plt.plot(list(inertias.keys()), list(inertias.values()))
plt.title('K-Means, Elbow Method')
plt.xlabel("Number of clusters, K")
plt.ylabel("Inertia")
plt.figure()
plt.grid(True)
plt.plot(list(silhouettes.keys()), list(silhouettes.values()))
plt.title('K-Means, Elbow Method')
plt.xlabel("Number of clusters, K")
plt.ylabel("Silhouette")
print("Clustering sparse data with %s" % km)
t0 = time()
km.fit(X)
print("done in %0.3fs" % (time() - t0))
print()
print("Homogeneity: %0.3f" % metrics.homogeneity_score(labels, km.labels_))
print("Completeness: %0.3f" % metrics.completeness_score(labels, km.labels_))
print("V-measure: %0.3f" % metrics.v_measure_score(labels, km.labels_))
print(
"Adjusted Rand-Index: %.3f"
% metrics.adjusted_rand_score(labels, km.labels_)
)
print(
"Silhouette Coefficient: %0.3f"
% metrics.silhouette_score(X, km.labels_, sample_size=1000)
)
print()
if not opts.use_hashing:
print("Top terms per cluster:")
if opts.n_components:
original_space_centroids = svd.inverse_transform(km.cluster_centers_)
order_centroids = original_space_centroids.argsort()[:, ::-1]
else:
order_centroids = km.cluster_centers_.argsort()[:, ::-1]
terms = vectorizer.get_feature_names()