def initialize(method, k, G, X, W):
if method == "spectral":
z0 = init.topeigen(k, G, W)
elif method == "k-means++":
z0 = init.kmeans_plus(k, X)
else:
z0 = np.random.randint(0, k, len(X))
Z0 = eclust.ztoZ(z0)
return Z0
def initialize(method, k, G, X, W):
if method == "spectral":
z0 = init.topeigen(k, G, W)
elif method == "k-means++":
z0 = init.kmeans_plus(k, X)
else:
z0 = np.random.randint(0, k, len(X))
Z0 = eclust.ztoZ(z0)
return Z0
def spectral(k, X, G, W=None, run_times=5):
if type(W) == type(None):
W = np.eye(len(X))
best_score = -np.inf
for _ in range(run_times):
zh = init.topeigen(k, G, W, run_times=run_times)
Zh = eclust.ztoZ(zh)
score = eclust.objective(Zh, G, W)
if score > best_score:
best_score = score
best_z = zh
return best_z
def spectral(k, X, G, W=None, run_times=5):
if type(W) == type(None):
W = np.eye(len(X))
best_score = -np.inf
for _ in range(run_times):
zh = init.topeigen(k, G, W, run_times=run_times)
Zh = eclust.ztoZ(zh)
score = eclust.objective(Zh, G, W)
if score > best_score:
best_score = score
best_z = zh
return best_z
def kernel_kgroups(k, X, G, W=None, run_times=5, ini="k-means++"):
if type(W) == type(None):
W = np.eye(len(X))
best_score = -np.inf
for _ in range(run_times):
Z0 = initialize(ini, k, G, X, W)
zh = eclust.kernel_kgroups(k, G, Z0, W, max_iter=300)
Zh = eclust.ztoZ(zh)
score = eclust.objective(Zh, G, W)
if score > best_score:
best_score = score
best_z = zh
return best_z
def kernel_kgroups(k, X, G, W=None, run_times=5, ini="k-means++"):
if type(W) == type(None):
W = np.eye(len(X))
best_score = -np.inf
for _ in range(run_times):
Z0 = initialize(ini, k, G, X, W)
zh = eclust.kernel_kgroups(k, G, Z0, W, max_iter=300)
Zh = eclust.ztoZ(zh)
score = eclust.objective(Zh, G, W)
if score > best_score:
best_score = score
best_z = zh
return best_z
#r.append(wrapper.kernel_kmeans(6, data, G, run_times=10, ini='random'))
r.append(wrapper.kernel_kmeans(6, data, G, run_times=10, ini='k-means++'))
#r.append(wrapper.kernel_kmeans(6, data, G, run_times=10, ini='spectral'))
#r.append(wrapper.kernel_kgroups(6,data,G,run_times=10, ini='random'))
r.append(wrapper.kernel_kgroups(6,data,G,run_times=10, ini='k-means++'))
#r.append(wrapper.kernel_kgroups(6,data,G,run_times=10, ini='spectral'))
t = PrettyTable(['Algorithm', 'Accuracy', 'A-Rand'])
algos = ['kmeans', 'GMM', 'spectral clustering', 'spectral',
'kernel k-means', 'kernel k-groups']
for algo, zh in zip(algos, r):
t.add_row([algo,
metric.accuracy(z, zh),
sklearn.metrics.adjusted_rand_score(z, zh)
])
print t
Z = np.array(eclust.ztoZ(z), dtype=int)
Zh = np.array(eclust.ztoZ(zh), dtype=int)
df = pd.DataFrame(Z)
df.to_csv('data/dermatology_true_label_matrix.csv', index=False, header=None)
df = pd.DataFrame(Zh)
df.to_csv('data/dermatology_pred_label_matrix.csv', index=False, header=None)
num_experiments = 10
table = np.zeros((num_experiments, 5))
for i in range(num_experiments):
X, z = data.univariate_lognormal([0, -1.5], [0.3, 1.5], [100, 100])
#X, z = data.univariate_normal([0, 5], [1, 22], [15, 15])
Y = np.array([[x] for x in X])
k = 2
# 1D energy clustering
zh, cost = two_clusters1D(X)
table[i,0] = accuracy(z, zh)
# initialization
z0 = initialization.kmeanspp(k, Y, ret='labels')
Z0 = eclust.ztoZ(z0)
rho = lambda x, y: np.linalg.norm(x-y)
G = eclust.kernel_matrix(Y, rho)
z1 = initialization.spectral(k, G)
Z1 = eclust.ztoZ(z1)
# Hartigan's method
zh = eclust.energy_hartigan(k, G, Z0)
table[i,1] = accuracy(z, zh)
zh = eclust.energy_hartigan(k, G, Z1)
table[i,2] = accuracy(z, zh)
# standard k-means
km = KMeans(2)
zh = km.fit_predict(Y)
r.append(wrapper.kernel_kmeans(6, data, G, run_times=10, ini='k-means++'))
#r.append(wrapper.kernel_kmeans(6, data, G, run_times=10, ini='spectral'))
#r.append(wrapper.kernel_kgroups(6,data,G,run_times=10, ini='random'))
r.append(wrapper.kernel_kgroups(6, data, G, run_times=10, ini='k-means++'))
#r.append(wrapper.kernel_kgroups(6,data,G,run_times=10, ini='spectral'))
t = PrettyTable(['Algorithm', 'Accuracy', 'A-Rand'])
algos = [
'kmeans', 'GMM', 'spectral clustering', 'spectral', 'kernel k-means',
'kernel k-groups'
]
for algo, zh in zip(algos, r):
t.add_row([
algo,
metric.accuracy(z, zh),
sklearn.metrics.adjusted_rand_score(z, zh)
])
print t
Z = np.array(eclust.ztoZ(z), dtype=int)
Zh = np.array(eclust.ztoZ(zh), dtype=int)
df = pd.DataFrame(Z)
df.to_csv('data/dermatology_true_label_matrix.csv', index=False, header=None)
df = pd.DataFrame(Zh)
df.to_csv('data/dermatology_pred_label_matrix.csv', index=False, header=None)